Welcome to FindFace Multi documentation!
FindFace Multi is a multifunctional multi-object video analytics software, based on FindFace Enterprise Server, a cutting-edge AI recognition technology. FindFace Multi is a turnkey solution that you can harness in such areas as retail, banking, social networking, entertainment, sports, event management, dating services, video surveillance, public safety, homeland security, and others.
FindFace Multi can detect, identify, and analyze the following objects in the video:
Human faces, along with recognition of such facial attributes as gender, age, emotions, glasses, face mask, beard, and many others. The integrated 2D anti-spoofing system ensures that it is a live person in front of a camera and eliminates the possibility of fraud using images on paper or mobile device screens.
Human bodies (silhouettes), along with recognition of clothing type and color.
Vehicles, with recognition of such vehicle attributes as make, model, body style, color, license plate number, and others.
After FindFace Multi identifies and analyzes an object, it notifies responsible officials about its appearance within fractions of a second. Additional information about the object, such as a person’s gender, age, car license plate number, etc., is displayed in the relevant recognition event.
FindFace Multi supports the integration of third-party solutions via HTTP API and webhooks, so you can enhance your current system or application with multi-object recognition functionality.
Core features
AI-based platform.
Comprehensive records that can accommodate aggregated information about a person and a car/vehicle.
Fast and robust real-time video monitoring against a record index.
Multi-object identification and analytics: faces, bodies (silhouettes), vehicles.
Support for live video and archives, most video formats and codecs that FFmpeg can decode.
Advanced camera management, including ONVIF support, the possibility of changing video orientation, fine-tuning a camera for each object type, video recording (if enabled).
Multi-object verification: faces, bodies, vehicles.
AI recognition of gender, age, emotions, glasses, beard, face mask, and other face attributes.
AI face liveness detector.
AI recognition of an exact person and vehicle.
AI recognition of clothing type and color.
AI recognition of special vehicles, a vehicle’s make, model, body style, color, and license plate number.
Database search for faces, bodies, vehicles.
Possibility of counting faces, bodies, and vehicles on connected cameras and measuring distance between bodies. Single- and multi-camera counting support.
Video surveillance. Video Recorder.
Possibility of monitoring the presence of people in given areas, using rules and monitoring schedules.
Automatic clustering of objects of the same origin (face/body images belonging to the same person, images of the same vehicle), enriched with the end-to-end integration with the record index.
Deployment environment
Developer-friendly installer and user-friendly interface.
Single- and multi-host deployment.
Increased performance and fault-tolerance in high load systems with numerous cameras and clients.
Network or on-premise licensing.
CPU- and GPU-based acceleration for your choice.
System security
Advanced user management.
Authentication based on a password, certificate, and face recognition for guaranteed system protection.
Record security.
Comprehensive, friendly, searchable audit logs.
Backup and recovery utilities.
Possibility of monitoring user sessions and blocking devices without deactivating user accounts.
Ethical data usage
Full compliance with personal data protection laws (GDPR and similar).
Make the most of your system
Social interaction analysis.
Audience analysis.
Detailed reports on face recognition events, episodes, search events, clusters, counters, cameras, records, audit logs, and audience.
Face liveness detector as a standalone service.
Useful little things
Quick record index creation.
Complete record customization.
Deduplication support for events and records.
Extended set of search filters.
Scheduled database cleanup.
Integration
Integration via HTTP API and webhooks.
Integrations with favored vendors.
Edge device integration.
What’s New in FindFace Multi 2.1
New Features:
Line Crossing: monitor the intersection of a predetermined virtual line by an individual or a vehicle.
Alarm Monitor: get notifications about alarming events into a separate desktop application or its web version.
Analytical BI Service: data visualization and exploration tool.
Router URL in camera settings: send detects to the external processing units.
Counter health monitoring: filter counters to find out the ones with technical issues.
Technical Changes:
Integration with Annex.
See Alarm Monitor
Tiered Data Storage: hot-cold storage to smooth out system peaks.
Enhanced Algorithms, UI, UX:
Webhooks configurator: in the web interface, create and configure a webhook, either in a visual or code mode.
VMS cleanup improved functionality.
See:
Enrich a record with a photo of the detected object directly from the event.
New and updated neural networks with enhanced characteristics.
Fresh Neural Networks:
Vehicle detection: an updated version
car.gustav_accurate.006and a new modelcar.gustav_normal.004with better accuracy in vehicle detection.Body detection: brand-new body detection models
body.gustav_accurate.019with enhanced characteristics andbody.jasmine_fast_nomany.018(GPU only), used for body detection when there is a small number (10-20) of objects in a frame.Face recognition: brand-new
nectarine_xs_320andnectarine_m_160face recognition models, much faster and more accurate than preceding models.Vehicle recognition: a brand-new
bottasneural network model with better accuracy in vehicle recognition.Body recognition: a brand-new
pedrec.durgabody recognition model with significantly improved quality of recognition for various retail scenarios (stores and malls).Face attribute recognition: an updated version
headpose.v3, more accurate and faster.
System Administrator’s Guide
This chapter is all about FindFace deployment and further updates and maintenance during exploitation.
Architecture
Though you mostly interact with FindFace Multi through its web interface, be sure to take a minute to learn the FindFace Multi architecture. This knowledge is essential for the FindFace Multi deployment, integration, maintenance, and troubleshooting.
In this chapter:
Recognition Objects and Recognition Process
FindFace Multi can recognize the following objects and their attributes:
human faces
human bodies (silhouettes)
vehicles
FindFace Multi detects an object in the photo or video and prepares its image through normalization. The normalized image is then used for extracting the object’s feature vector (an n-dimensional vector of numerical features that represent the object). Object feature vectors are stored in the database and further used for verification and identification purposes.
Docker Platform
FindFace Multi is deployed in Docker, a set of platform as a service (PaaS) products that use OS-level virtualization to deliver software in packages called containers. Each FindFace Multi service runs in a Docker container.
Architectural Elements
FindFace Multi consists of the following fundamental architectural elements:
FindFace core, a cutting-edge AI-based recognition technology that can be used as a separate product FindFace Enterprise Server.
FindFace Multi, which is a turnkey application module for FindFace Enterprise Server.
The FindFace core internal architecture allows for embedding Video Recorder, an additional functionality that records, stores, and plays back video data from cameras.
Architecture scheme
FindFace core
There is a separate product FindFace Enterprise Server inside the FindFace core. It includes the following components:
Component |
Ports in use |
Description |
Vendor |
|---|---|---|---|
findface-extraction-api |
18666 |
Service that uses neural networks to detect an object in an image and extract its feature vector. It also recognizes object attributes (for example, gender, age, emotions, beard, glasses, face mask - for face objects). CPU- or GPU-acceleration. |
NtechLab own deployment |
findface-sf-api |
18411 |
Service that implements the internal HTTP API for object detection and recognition. |
|
findface-tarantool-server |
32001, shard ports (default 330xx, 81xx) |
Service that provides interaction between the |
|
findface-upload |
3333 |
NginX-based web server used as a storage for original images, thumbnails, and
normalized object images. If Video Recorder is installed, |
|
findface-facerouter |
18820 |
Service used to define processing directives for detected objects. In FindFace
Multi, its functions are performed by |
|
findface-video-manager |
18810, 18811 |
Service, part of the video object detection module, that is used for managing the video object detection functionality, configuring the video object detector settings and specifying the list of to-be-processed video streams. |
|
findface-video-worker |
18999 |
Service, part of the video object detection module, that recognizes an object in
the video and posts its normalized image, full frame and metadata (such as
detection time) to the |
|
findface-ntls |
443 (TCP), 3133, 3185 |
License server that interfaces with the NtechLab Global License Server, a USB dongle, or hardware fingerprint to verify the license of your FindFace Multi instance. |
|
findface-counter |
18300 |
Service used for event deduplication. |
|
findface-liveness-api |
18301 |
Besides the embedded functionality provided by |
|
Tarantool |
Shard ports (default 330xx, 81xx) |
Third-party software that implements the feature vector database that stores extracted object feature vectors and identification events. The system data, records, user accounts, and camera settings are stored in PostgreSQL (part of the FindFace Multi application module). |
|
etcd |
2379 |
Third-party software that implements a distributed key-value store for
|
|
NginX |
80; SSL: 8002, 8003, 443, 80 |
Third-party software that implements the system web interfaces. |
|
memcached |
11211 |
Third-party software that implements a distributed memory caching system.
Used by |
FindFace Multi Application Module
The FindFace Multi application module includes the following components:
Component |
Ports in use |
Description |
Vendor |
|---|---|---|---|
findface-multi-legacy |
Configurable |
Service that serves as a gateway to the FindFace core. Provides interaction between the FindFace core and the web interface, the system functioning as a whole, HTTP and web socket, object monitoring, event notifications, episodes, webhooks, object clusterization, and counters. |
NtechLab own deployment |
findface-multi-pause |
n/a |
Internal services that assist |
|
findface-multi-audit |
8012, 8013, 8014 |
||
findface-multi-identity-provider |
8022, 8023, 8024 |
||
findface-onvif-discovery |
n/a |
||
findface-multi-file-mover |
n/a |
A file moving service that is used for tiered data storage. The service transfers images from the short-term storage to the long-term storage. |
|
findface-multi-ui |
Configurable |
Main web interface used to interact with FindFace Multi. Based on the Django framework. Allows you to work with object recognition events, search for objects, manage cameras, users, records, watch lists, collect real-time statistics, and many more. |
|
NATS |
4222 |
Third-party software that implements a message broker in |
|
etcd |
2379 |
Third-party software that implements locks in the |
|
Pgbouncer |
5439 |
Third-party software, a lightweight connection pooler for PostgreSQL. Optional, used to increase the database performance under high load. |
|
PostgreSQL |
5432 |
Third-party software that implements the main system database. This database stores records of persons and vehicles and data for internal use, including user accounts and camera settings. The object feature vectors and object identification events are stored in Tarantool (part of the FindFace core). |
Video Recorder
Video Recorder is an optional part of the FindFace core. It includes the following services:
Component |
Ports in use |
Description |
Vendor |
|---|---|---|---|
findface-video-storage |
18611 |
Service that implements video chunk management. It takes video chunks from the
|
NtechLab own deployment |
findface-video-streamer |
9000 |
After receiving a |
|
MongoDB |
27017 |
Third-party software that implements the Video Recorder database. The database stores
meta-information of the video chunks, including their location. The video chunks
themselves are stored in the |
See also
Single- and Multi-Host Deployment
You can deploy FindFace Multi on a single host or in a multi-host environment. If you opt for the latter, we offer you one of the following deployment schemes:
Deploy FindFace Multi on a principal host and distribute additional
findface-video-workerinstances across remote hosts.Distribute the FindFace Multi services across multiple hosts. If necessary, set up load balancing. Contact our experts by support@ntechlab.com for guidance.
CPU- and GPU-acceleration
The findface-extraction-api and findface-video-worker services can be either CPU- or GPU-based. You will have an opportunity to choose the acceleration type you need during the installation.
Important
Refer to Requirements when choosing hardware configuration.
Important
If the resolution of a camera(s) in use is more than 1280x720px, it is strongly recommended to use the GPU-accelerated package findface-video-worker.
Note
The liveness detector is much slower on CPU than on GPU.
Requirements
In this chapter:
System Requirements for Basic Configuration
To calculate the FindFace host(s) characteristics, use the requirements provided below.
Tip
Be sure to learn about the FindFace architecture first.
Important
If the video resolution is more than 1280x720px, it is strongly recommended to use the GPU-accelerated package findface-video-worker-gpu.
Important
On AMD CPU servers, the full functionality of the CPU-accelerated findface-extraction-api service is not guaranteed. Use the GPU-accelerated service findface-extraction-api-gpu along with the GPU-version of neural networks instead.
Note
In the case of a high-load system (~> 15 events per second), be sure to use an SSD.
Minimum |
Recommended |
|
|---|---|---|
CPU |
Intel Core i5 CPU with 4+ physical cores 3+ GHz. AVX2 support |
Intel Xeon Silver/Gold with 6+ physical cores |
The own needs of FindFace require 2 cores HT > 2.5 GHz. The characteristics also depend on the number of
video files in process. A single video file |
||
GPU (optional) |
GeForce® RTX 3060 12 GB |
NVIDIA A10 |
Supported devices: NVIDIA, Pascal architecture and above. Note: NVIDIA GeForce RTX 40 Series graphics cards are currently not supported. |
||
RAM |
16 Gb |
32+ Gb |
The RAM consumption depends on:
A single video file |
||
HDD (SSD for best performance) |
65 Gb |
65+ Gb |
The own needs of FindFace require 65 GB. The total volume is subject to the required depth of the event archive in the database and in the log, at the rate of 1.5 Mb per 1 event. |
||
Operating system |
Ubuntu Server / Desktop from 18 to 22, x64 only, RHEL, CentOS 7, Debian 11. |
|
Note
You can also use an Intel-based VM if there is AVX2 support, and eight physical cores are allocated exclusively to the VM.
Tip
For more accurate hardware selection, contact our support team by support@ntechlab.com.
Required Administrator Skills
A FindFace Multi administrator must know and understand OS, on which the product instance is deployed, at the level of an advanced user.
Video File Formats
FindFace Multi supports a wide variety of file formats depending on the acceleration type, CPU or GPU.
Both CPU- and GPU-accelerated instances support all FFmpeg codecs. In addition to that, the following codecs are supported:
CPU-based acceleration: FLV (both as a codec and as a container), H263, H264, H265, MJPEG, VP8, VP9, MPEG1VIDEO, MPEG2VIDEO, MSMPEG4v2, MSMPEG4v3.
GPU-based acceleration: MJPEG, H264, H265, VP9, and others, depending on the list of codecs supported by the used video card. Apart from that, for instances with
video-worker-gpu, you can expand the number of supported codecs by enabling video decoding on the CPU, which is not available by default.
To enable video decoding on the CPU for GPU-based acceleration, do the following:
Open the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlfile.sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml
Set
cpu: truein thevideo_decodersection.... video_decoder: cpu: true ...
Restart
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
Requirements for CCTV Cameras
Face Recognition
The primary requirements for installation and characteristics of CCTV cameras in your FindFace Multi-based face recognition system are the following:
For correct face detection in a video stream, mount the camera so that the face of each individual entering the monitored area surely appears in the camera field of view.
The vertical tilt angle of the camera should not exceed 15°. The vertical tilt is a deviation of the camera’s optical axis from the horizontal plane, positioned at the face center’s level for an average height person (160 cm).
The horizontal deflection angle should not exceed 30°. The horizontal deflection is a deviation of the camera’s optical axis from the motion vector of the main flow of objects subject to recognition.
The minimum pixel density required for identification is 500 pixels/m (roughly corresponds to a face width of 80 pixels).
Select such a focal length of the camera’s lenses that provides the required pixel density at a predetermined distance to the recognition objects. The picture below demonstrates how to calculate the focal length subject to the distance between the camera and recognition objects. Estimating the focal length for a particular camera requires either calculators or a methodology provided by the camera manufacturer.
The exposure must be adjusted so that the face images are sharp (“in focus”), non-blurred, and evenly lit (not overlit or too dark).
For imperfect lighting conditions such as flare, too bright or too dim illumination, choose cameras with WDR hardware (Wide Dynamic Range) or other technologies that provide compensation for backlight and low illumination. Consider BLC, HLC, DNR, high optical sensitivity, Smart infrared backlight, AGC, and such.
Video compression: most video formats and codecs that FFmpeg can decode.
Video stream delivery protocols: RTSP, HTTP.
Tip
To calculate the precise hardware configuration tailored to your purposes, contact our experts by support@ntechlab.com.
Body and Vehicle Recognition
Please contact our technical support team (support@ntechlab.com) to get requirements for installation and characteristics of CCTV cameras for body and vehicle recognition.
Deploy and Remove FindFace Multi
Docker platform
FindFace Multi 2.0+ is Docker-based. You must install and start a set of Docker products before proceeding with the FindFace Multi deployment. For your convenience, this chapter contains the Ubuntu Server Preparation section covering the intricacies of installing Docker on Ubuntu. For other platforms, please refer to the Docker documentation.
NVIDIA driver and NVIDIA Container Runtime (GPU only)
If you intend to deploy FindFace Multi with GPU-acceleration, you need to install the NVIDIA driver and NVIDIA Container Runtime. Again, you will find the relevant information in the Ubuntu Server Preparation section.
Deployment options
After you are finished with the server preparation, you are all set for the FindFace Multi deployment. You are provided with the following options here:
Automatically install FindFace Multi standalone in one run. Being the simplest, this installation type is excellent to kick-start your experience with FindFace Multi. We recommend choosing it if you are a first-timer. See FindFace Multi Standalone Automated Deployment for guidance.
Automatically install FindFace Multi and configure it to interact with additional remote
findface-video-workerinstances. This installation type is for those who are deploying FindFace Multi in a multi-host environment. It requires a bit of technical expertise and knowledge of the product architecture. To automatically install only FindFace Video Worker (findface-video-worker) on a remote host, refer to the FindFace Video Worker Deployment on Remote Hosts section.Fully customized installation. It requires fundamental understanding of the product architecture. See Fully Customized Installation.
Note
If you select the installation type #3, keep in mind to install necessary neural network models along with the findface-extraction-api component.
Installer questions and automatic deployment from the installation file
Before starting the actual installation, the installer asks you a few questions and performs several automated checks to ensure that the host meets the system requirements. After filling out each prompt, press Enter.
To install the same configuration of FindFace Multi on a different host, use the automatic deployment from the installation file. In this case, you won’t have to answer the installation questions again. The exact path to the installation file is displayed right after the last question, before the installer starts off active progress:
[I 2023-02-09 11:13:37,187 main:142] Your answers were saved to /tmp/findface-installer-p01n9sn3.json
Important
Be sure to remove fields *.config, exp_ip, and int_ip before installing FindFace Multi on a host with a different IP address.
To launch the automatic deployment from the /tmp/<findface-installer-*>.json file, execute:
sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.json
Post-development procedures and how-to’s
Browse through the Post-deployment Procedures and How-to’s section to learn how to set the time zone, license your instance, and configure logging. This section will also provide you with few basic commands that will help you to kick-start your work with FindFace Multi containers, in case you are a newbie with Docker.
Important
Starting the GPU-accelerated services findface-extraction-api and findface-video-worker for the first time after deployment may take a considerable amount of time due to the caching process (up to 45 minutes).
Important
Although FindFace Multi provides tools to ensure its protection from unauthorized access, they are not replacing a properly configured firewall. Be sure to use a firewall to heighten the FindFace Multi network protection.
Instance removal
To remove your instance, you must run a set of commands. See the Remove FindFace Multi Instance section.
Ubuntu Server Preparation
To prepare a server on Ubuntu for the FindFace Multi deployment, follow the instructions below minding the sequence.
Note
For other platforms, please refer to the following resources:
In this section:
GPU: Install NVIDIA Drivers
The first step of the server preparation is the NVIDIA driver installation. It’s applicable only for the GPU configuration. Go straight to the Docker installation if your configuration is CPU-accelerated.
With the GPU-accelerated FindFace Multi, you’ll need the NVIDIA driver version 530 or above. Add the NVIDIA repository and install an applicable driver from it.
Warning
We do not recommend using a .run installer from NVIDIA Driver Downloads instead, as its drivers may conflict with the drivers installed by packages.
To install the 530 driver from a repository, do the following:
Install the repository signature key:
arch=$(uname -m); version=$(. /etc/os-release; echo $ID$VERSION_ID | sed -r 's/\.//g'); sudo bash -c \ "sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/$ID$version/$arch/3bf863cc.pub \ && apt update"
Install
aptitude:sudo apt-get install aptitude
Install
nvidia-driver-530:sudo aptitude install nvidia-driver-530
Reboot the system:
sudo reboot
Install Docker Products
Docker products must be installed on both CPU and GPU servers. Do the following:
Update the
aptpackage index and install packages to allowaptto use a repository over HTTPS.sudo apt-get update sudo apt-get install \ ca-certificates \ curl \ gnupg \ lsb-release
Tip
You may receive the following errors when running
sudo apt-get install \:E: Could not get lock /var/lib/dpkg/lock-frontend - open (11: Resource temporarily unavailable) E: Unable to acquire the dpkg frontend lock (/var/lib/dpkg/lock-frontend), is another process using it?
There are two ways to resolve them:
Forcibly terminate all the
apt-getprocesses currently running in the system.sudo killall apt apt-get
If the previous command didn’t help, run the set of commands below. It’s fine if some of the to-be-deleted directories do not exist — just keep going.
sudo rm /var/lib/apt/lists/lock sudo rm /var/cache/apt/archives/lock sudo rm /var/lib/dpkg/lock sudo rm /var/lib/dpkg/lock-frontend sudo dpkg --configure -a
Add the Docker’s official GPG key (GNU Privacy Guard key) to the host.
sudo mkdir -m 0755 -p /etc/apt/keyrings curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg
Set up the Docker repository.
echo \ "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu \ $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
Update the
aptpackage index one more time.sudo apt-get update
Tip
If you have received a GPG error when running this command, try granting read permission for the Docker public key file before updating the package index.
sudo chmod a+r /etc/apt/keyrings/docker.gpg sudo apt-get update
Install
24.*versions of Docker products.sudo apt-get install docker-ce=5:24* docker-ce-cli=5:24* containerd.io docker-buildx-plugin
Check whether the Docker installation was a success. The following command downloads a test image and runs it in a container. When the container runs, it prints a confirmation message and exits.
sudo docker run hello-world Unable to find image 'hello-world:latest' locally latest: Pulling from library/hello-world 2db29710123e: Pull complete Digest: sha256:aa0cc8055b82dc2509bed2e19b275c8f463506616377219d9642221ab53cf9fe Status: Downloaded newer image for hello-world:latest Hello from Docker! This message shows that your installation appears to be working correctly. To generate this message, Docker took the following steps: 1. The Docker client contacted the Docker daemon. 2. The Docker daemon pulled the "hello-world" image from the Docker Hub. (amd64) 3. The Docker daemon created a new container from that image which runs the executable that produces the output you are currently reading. 4. The Docker daemon streamed that output to the Docker client, which sent it to your terminal. To try something more ambitious, you can run an Ubuntu container with: $ docker run -it ubuntu bash Share images, automate workflows, and more with a free Docker ID: https://hub.docker.com/ For more examples and ideas, visit: https://docs.docker.com/get-started/
Install
docker-compose.sudo curl -SL https://github.com/docker/compose/releases/download/v2.15.1/docker-compose-linux-x86_64 -o /usr/local/bin/docker-compose sudo chmod +x /usr/local/bin/docker-compose sudo ln -s /usr/local/bin/docker-compose /usr/bin/docker-compose
Perform the Docker Engine post-installation procedures to ease your future work with Docker and FindFace Multi containers. Once you can manage Docker as a non-root user, you don’t have to apply
sudoin the commands related to Docker.sudo groupadd docker sudo usermod -aG docker $USER newgrp docker
Configure the Docker network.
sudo su BIP=10.$((RANDOM % 256)).$((RANDOM % 256)).1 cat > /etc/docker/daemon.json <<EOF { "bip": "$BIP/24", "fixed-cidr": "$BIP/24" } EOF
GPU: Install NVIDIA Container Runtime
To deploy containerized GPU-accelerated FindFace Multi, you need NVIDIA Container Runtime. We recommend installing NVIDIA Container Toolkit that includes this runtime. Do the following:
Specify the repository and install NVIDIA Container Toolkit from it by executing the following commands.
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \ && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \ && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \ sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \ sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list sudo apt-get update sudo apt-get install -y nvidia-container-toolkit sudo nvidia-ctk runtime configure --runtime=docker sudo systemctl restart dockerConfigure the Docker network. Configure usage of the NVIDIA Container Runtime installed along with NVIDIA Container Toolkit.
sudo su BIP=10.$((RANDOM % 256)).$((RANDOM % 256)) cat > /etc/docker/daemon.json <<EOF { "default-address-pools": [ {"base":"$BIP.0/16","size":24} ], "bip": "$BIP.1/24", "fixed-cidr": "$BIP.0/24", "runtimes": { "nvidia": { "path": "nvidia-container-runtime", "runtimeArgs": [] } }, "default-runtime": "nvidia" } EOFRestart Docker.
systemctl restart docker
Now you are all set to install FindFace Multi. Refer to the following sections:
CentOS 7 Server Preparation
To prepare a server on CentOS 7 for the FindFace Multi deployment, follow the instructions below minding the sequence.
Note
For other platforms, please refer to the following resources:
In this section:
Install Updates
Run updates and reboot the server.
sudo yum update sudo reboot
Install
fuseby running the following command.sudo yum -y install fuse
GPU: Install NVIDIA Drivers
The first step of the server preparation is the NVIDIA driver installation. It’s applicable only for the GPU configuration. Go straight to the Docker installation if your configuration is CPU-accelerated.
With the GPU-accelerated FindFace Multi, you’ll need the NVIDIA driver version 530 or above. Download a relevant .run installer from NVIDIA Driver Downloads.
Since you use a .run installer from NVIDIA, the following dependencies must be installed:
sudo yum install kernel-devel gcc kernel-headers
Install Docker Products
Docker products must be installed on both CPU and GPU servers. Do the following:
Install the
yum-utilspackage (which provides theyum-config-managerutility) and set up the repository.sudo yum install -y yum-utils sudo yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo
Install
24.*versions of Docker products.sudo yum install docker-ce-3:24* docker-ce-cli-1:24* docker-ce-rootless-extras-24* containerd.io docker-buildx-plugin docker-compose-plugin
Start and enable Docker.
sudo systemctl start docker sudo systemctl enable docker
Check whether the Docker installation was a success. The following command downloads a test image and runs it in a container. When the container runs, it prints a confirmation message and exits.
sudo docker run hello-world Unable to find image 'hello-world:latest' locally latest: Pulling from library/hello-world 2db29710123e: Pull complete Digest: sha256:aa0cc8055b82dc2509bed2e19b275c8f463506616377219d9642221ab53cf9fe Status: Downloaded newer image for hello-world:latest Hello from Docker! This message shows that your installation appears to be working correctly. To generate this message, Docker took the following steps: 1. The Docker client contacted the Docker daemon. 2. The Docker daemon pulled the "hello-world" image from the Docker Hub. (amd64) 3. The Docker daemon created a new container from that image which runs the executable that produces the output you are currently reading. 4. The Docker daemon streamed that output to the Docker client, which sent it to your terminal. To try something more ambitious, you can run an Ubuntu container with: $ docker run -it ubuntu bash Share images, automate workflows, and more with a free Docker ID: https://hub.docker.com/ For more examples and ideas, visit: https://docs.docker.com/get-started/
Install
docker-compose.sudo curl -SL https://github.com/docker/compose/releases/download/v2.15.1/docker-compose-linux-x86_64 -o /usr/local/bin/docker-compose sudo chmod +x /usr/local/bin/docker-compose sudo ln -s /usr/local/bin/docker-compose /usr/bin/docker-compose
Perform the Docker Engine post-installation procedures to ease your future work with Docker and FindFace Multi containers. Once you can manage Docker as a non-root user, you don’t have to apply
sudoin the commands related to Docker.sudo groupadd docker sudo usermod -aG docker $USER newgrp docker
Configure the Docker network and
devicemapperusage.sudo su BIP=10.$((RANDOM % 256)).$((RANDOM % 256)).1 cat > /etc/docker/daemon.json <<EOF { "bip": "$BIP/24", "fixed-cidr": "$BIP/24", "storage-driver": "devicemapper" } EOF
GPU: Install NVIDIA Container Runtime
To deploy containerized GPU-accelerated FindFace Multi, you need NVIDIA Container Runtime. We recommend installing NVIDIA Container Toolkit that includes this runtime. Do the following:
Specify the repository and install NVIDIA Container Toolkit from it by executing the following commands.
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \ && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.repo | sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo sudo yum clean expire-cache sudo yum install -y nvidia-container-toolkit sudo nvidia-ctk runtime configure --runtime=docker sudo systemctl restart docker
Configure the Docker network,
devicemapperusage, and usage of the NVIDIA Container Runtime installed along with NVIDIA Container Toolkit.sudo su BIP=10.$((RANDOM % 256)).$((RANDOM % 256)) cat > /etc/docker/daemon.json <<EOF { "default-address-pools": [ {"base":"$BIP.0/16","size":24} ], "bip": "$BIP.1/24", "fixed-cidr": "$BIP.0/24", "runtimes": { "nvidia": { "path": "nvidia-container-runtime", "runtimeArgs": [] } }, "default-runtime": "nvidia", "storage-driver": "devicemapper" } EOFRestart Docker.
systemctl restart docker
Now you are all set to install FindFace Multi. Refer to the following sections:
Debian 11 Server Preparation
To prepare a server on Debian 11 for the FindFace Multi deployment, follow the instructions below minding the sequence.
Note
For other platforms, please refer to the following resources:
In this section:
Install FUSE
Install FUSE (Filesystem in Userspace) by running the following command.
sudo apt install fuse -y
GPU: Install NVIDIA Drivers
The first step of the server preparation is the NVIDIA driver installation. It’s applicable only for the GPU configuration. Go straight to the Docker installation if your configuration is CPU-accelerated.
With the GPU-accelerated FindFace Multi, you’ll need the NVIDIA driver version 530 or above. Download a relevant .run installer from NVIDIA Driver Downloads.
Since you use a .run installer from NVIDIA, the following dependencies must be installed:
sudo apt install linux-headers-$(uname -r)
sudo apt install build-essential
Install Docker Products
Docker products must be installed on both CPU and GPU servers. Do the following:
Update the
aptpackage index and install packages to allowaptto use a repository over HTTPS.sudo apt-get update sudo apt-get install ca-certificates curl gnupg
Add the Docker’s official GPG key (GNU Privacy Guard key) to the host.
sudo install -m 0755 -d /etc/apt/keyrings curl -fsSL https://download.docker.com/linux/debian/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg sudo chmod a+r /etc/apt/keyrings/docker.gpg
Set up the Docker repository.
echo \ "deb [arch="$(dpkg --print-architecture)" signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/debian \ "$(. /etc/os-release && echo "$VERSION_CODENAME")" stable" | \ sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
Update the
aptpackage index one more time.sudo apt-get update
Install
24.*versions of Docker products.sudo apt-get install docker-ce=5:24* docker-ce-cli=5:24* containerd.io docker-buildx-plugin
Check whether the Docker installation was a success.
sudo docker run hello-world
This command downloads a test image and runs it in a container. When the container runs, it prints a confirmation message and exits.
Install
docker-compose.sudo curl -SL https://github.com/docker/compose/releases/download/v2.15.1/docker-compose-linux-x86_64 -o /usr/local/bin/docker-compose sudo chmod +x /usr/local/bin/docker-compose sudo ln -s /usr/local/bin/docker-compose /usr/bin/docker-compose
Perform the Docker Engine post-installation procedures to ease your future work with Docker and FindFace Multi containers. Once you can manage Docker as a non-root user, you don’t have to apply
sudoin the commands related to Docker.sudo groupadd docker sudo usermod -aG docker $USER newgrp docker
Configure the Docker network.
sudo su BIP=10.$((RANDOM % 256)).$((RANDOM % 256)).1 cat > /etc/docker/daemon.json <<EOF { "bip": "$BIP/24", "fixed-cidr": "$BIP/24" } EOF
GPU: Install NVIDIA Container Runtime
To deploy containerized GPU-accelerated FindFace Multi, you need NVIDIA Container Runtime. We recommend installing NVIDIA Container Toolkit that includes this runtime. Do the following:
Specify the repository and install NVIDIA Container Toolkit from it by executing the following commands.
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \ && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \ && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \ sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \ sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list sudo apt-get update sudo apt-get install -y nvidia-container-toolkit sudo nvidia-ctk runtime configure --runtime=docker sudo systemctl restart dockerSwitch to the superuser account.
sudo su
Configure the Docker network. Configure usage of the NVIDIA Container Runtime installed along with NVIDIA Container Toolkit.
BIP=10.$((RANDOM % 256)).$((RANDOM % 256)) cat > /etc/docker/daemon.json <<EOF { "default-address-pools": [ {"base":"$BIP.0/16","size":24} ], "bip": "$BIP.1/24", "fixed-cidr": "$BIP.0/24", "runtimes": { "nvidia": { "path": "nvidia-container-runtime", "runtimeArgs": [] } }, "default-runtime": "nvidia" } EOFRestart Docker.
systemctl restart docker
Now you are all set to install FindFace Multi. Refer to the following sections:
FindFace Multi Standalone Automated Deployment
To deploy FindFace Multi as a standalone server in one run, follow the instructions below. Being the simplest, this installation type is excellent to start off your work with FindFace Multi. Be sure to meet the system requirements and prepare the server first.
Important
The FindFace Multi host must have a static IP address in order to be running successfully. To make the IP address static, open the /etc/network/interfaces file and modify the current primary network interface entry as shown in the case study below. Be sure to substitute the suggested addresses with the actual ones, subject to your network specification.
sudo vi /etc/network/interfaces
iface eth0 inet static
address 192.168.112.144
netmask 255.255.255.0
gateway 192.168.112.254
dns-nameservers 192.168.112.254
Restart networking.
sudo service networking restart
Be sure to edit the etc/network/interfaces file with extreme care. Please refer to the Ubuntu guide on networking before proceeding.
Do the following:
Download the installer file
findface-*.run.Put the
.runfile into some directory on the designated host (for example,/home/username).From this directory, make the
.runfile executable.Note
Be sure to specify the actual file name instead of
findface-*.chmod +x findface-*.run
Execute the
.runfile.sudo ./findface-*.run
The installer will ask you a few questions and perform several automated checks to ensure that the host meets the system requirements. After filling out each prompt, press Enter. The questions and answers are the following:
Q:
Which product should be installed?A:
11. [multi ] FindFace Multi 2. [server ] FindFace Server 3. [video-worker] FindFace Video Worker (default: multi) product> 1
Q:
Please choose installation type:A:
1- 1 [stand-alone ] Single Server - 2 [multi-worker] Single Server, Multiple video workers - 3 [images ] Don't install anything, just load the images - 4 [custom ] Fully customized installation (default: stand-alone) type> 1
Q:
Directory to install into:A: Specify the FindFace Multi installation directory. By default, it’s
/opt/findface-multi. Press Enter to confirm it. Otherwise, specify the directory of your choice and press Enter.Directory to install into: (default: /opt/findface-multi) dest_dir>
Q:
Do you want to install Video Recorder?(y/n)A:
yorn, subject to your needs.Do you want to install Video Recorder?(y/n) install_video_recorder> n
Q:
Found X interface(s). Which one should we announce as our external address?A: Choose the interface that you are going to use as the instance IP address.
Found 1 interface(s). Which one should we announce as our external address? - 1 [lo ] 127.0.0.1 - 2 [ens3 ] 192.168.112.254 (default: 192.168.112.254) ext_ip.advertised> 2
Q:
Which variant of Video Worker should be installed?A: Specify the
findface-video-workerpackage type, CPU or GPU.Which variant of Video Worker should be installed? - 1 [cpu] CPU-based implementation, slower but doesn't require GPU - 2 [gpu] CUDA-based implementation of video detector, requires NVIDIA GPU (default: cpu) findface-video-worker.variant> 1
Q:
Which variant of Extraction API should be installed?A: Specify the
findface-extraction-apipackage type, CPU or GPU.Which variant of Extraction API should be installed? - 1 [cpu] CPU-only implementation, slower but doesn't require GPU - 2 [gpu] CUDA-based implementation, faster, requires NVIDIA GPU (supports both CPU and GPU models) (default: cpu) findface-extraction-api.variant> 1
Q:
Do you want to configure detectors and features right now?(y/n)A: We recommend you to install and configure the object detection and object attribute recognition functionality straightaway during the installation. Answer
yto initiate the process. You can skip it, though, by answeringnand perform the necessary steps later, following the instructions in the Enable Face and Face Attribute Recognition, Enable Vehicle and Vehicle Attribute Recognition, and Enable Body and Body Attribute Recognition sections.Do you want to configure detectors and features right now?(y/n) configure> y
Q:
Please select detectors to install:A: This question appears if you have requested installation and configuration of the object detection and object attribute recognition functions. In the multiple choice form, the face detector is checked by default. Enter a relevant number to select an unpicked detector and vice versa. For example, to add the body and vehicle detector to your configuration , enter
2 3. Enterdoneto proceed.Please select detectors to install: - 1 [v] Face - 2 [ ] Body - 3 [ ] Car Enter keyword to select matching choices or -keyword to clear selection. Enter "done" to save your selection and proceed to another step. detectors>2 3 - 1 [v] Face - 2 [v] Body - 3 [v] Car Enter keyword to select matching choices or -keyword to clear selection. Enter "done" to save your selection and proceed to another step. detectors> done
Q:
Enable liveness and attempt to continue installation?(y/n)A: This question appears if you have requested installation and configuration of the object detection and object attribute recognition functions. To install the embedded face liveness detector, enter
y. Enternotherwise.Enable liveness and attempt to continue installation?(y/n) enable_liveness> y
Q:
Please select face features to install:A: This question appears if you have requested installation and configuration of the face detection and face attribute recognition functions. By default, all face attributes are subject to installation. Answer
doneto confirm it. If some attribute is not necessary, you can enter the keyword (number) related to it. For example, enter7to exclude the head pose recognition. Then enterdone.Questions on body and vehicle attribute recognition will be analogous, with multiple choice forms depending on the requested detector type.
Please select face features to install: - 1 [v] Age - 2 [v] Gender - 3 [v] Emotions - 4 [v] Beard - 5 [v] Glasses - 6 [v] Medicine masks - 7 [v] Headpose Enter keyword to select matching choices or -keyword to clear selection. Enter "done" to save your selection and proceed to another step. face_features> done
Questions on body and vehicle attribute recognition will be analogous, with multiple choice forms depending on the requested detector type.
Q:
Please set findface-multi admin passwordA: Set the Super Administrator (
superuser) password.Please set findface-multi admin password findface-multi-admin-password> admin
Q:
Do you want to install Analytics service?(y/n)A: By answering
yyou will install the service, which will process FindFace Multi data for further use in analytical purposes. The analytics works for line crossing only at the moment.Do you want to install Analytics service?(y/n) enable_analytics> y
Q:
Do you want to install Superset BI?(y/n)A: Installation of the external BI system will allow you to view FindFace Multi data (currently available for line crossing only) in the external system in the form of a graph. FindFace Multi serves as a data source, the external BI system helps visualize data. By answering
yyou choose to install the Apache Superset open-source software application. You will be able to log in to the system with the default credentialsadmin | admin; a password can further be changed in a system web-interface. If you prefer your own BI system, you can skip the suggested installation and answern.Do you want to install Superset BI?(y/n) install_superset> y
After the final question you will see a path to a JSON file with your answers:
Your answers were saved to /tmp/findface-installer-t8qk_isw.json
The installer will pull the FindFace Multi images from the Ntechlab registry and start the following services in Docker containers:
Service
Container
Configuration
findface-multi-pause
findface-multi-findface-multi-pause-1
Started
nats-jetstream
findface-multi-nats-jetstream-1
Started
mongodb
findface-multi-mongodb-1
Installed with Video Recorder. Started
findface-ntls
findface-multi-findface-ntls-1
Started
nats
findface-multi-nats-1
Started
postgresql
findface-multi-postgresql-1
Started
memcached
findface-multi-memcached-1
Started
findface-upload
findface-multi-findface-upload-1
Started
etcd
findface-multi-etcd-1
Started
findface-sf-api
findface-multi-findface-sf-api-1
Started
findface-extraction-api
findface-multi-findface-extraction-api-1
Started (CPU/GPU-acceleration)
findface-tarantool-server-shard-*
findface-multi-findface-tarantool-server-shard-*-1
Started. The number of shards is calculated using the formula:
N = min(max(min(mem_mb // 2000, cpu_cores), 1), 16 * cpu_cores). I.e., it is equal to the RAM size in MB divided by 2000, or the number of CPU physical cores (but at least one shard), or the number of CPU physical cores multiplied by 16, if the first obtained value is greater.findface-video-manager
findface-multi-findface-video-manager-1
Started
findface-counter
findface-multi-findface-counter-1
Started
findface-video-streamer
findface-multi-findface-video-streamer-1
Installed with Video Recorder. Started
findface-video-storage
findface-multi-findface-video-storage-1
Installed with Video Recorder. Started
pgbouncer
findface-multi-pgbouncer-1
Started
findface-liveness-api
findface-multi-findface-liveness-api-1
Started
findface-video-worker
findface-multi-findface-video-worker-1
Started (CPU/GPU-acceleration)
findface-onvif-discovery
findface-multi-findface-onvif-discovery-1
Started
findface-multi-audit
findface-multi-findface-multi-audit-1
Started
findface-multi-identity-provider
findface-multi-findface-multi-identity-provider-1
Started
findface-multi-legacy
findface-multi-findface-multi-legacy-1
Started
findface-multi-ui
findface-multi-findface-multi-ui-1
Started
findface-multi-line-crossing-analytics
findface-multi-findface-multi-line-crossing-analytics-1
Started
redis
findface-multi-redis-1
Started
rabbitmq
findface-multi-rabbitmq-1
Started
timescaledb
findface-multi-timescaledb-1
Started
findface-image-crop
findface-multi-findface-image-crop-1
Started
backend_api
findface-multi-backend_api-1
Started
service_notifier_ws
findface-multi-service_notifier_ws-1
Started
service_notifier_tg
findface-multi-service_notifier_tg-1
Started
backend_provider
findface-multi-backend_provider-1
Started
service_alarmer
findface-multi-service_alarmer-1
Started
superset-worker
findface-multi-superset-worker-1
Started
superset-init
findface-multi-superset-init-1
Started
superset-worker-beat
findface-multi-superset-worker-beat-1
Started
superset
findface-multi-superset-1
Started
findface-multi-sync
findface-multi-findface-multi-sync-1
Started
alarm-web-app
findface-multi-alarm-app-1
Started
findface-multi-file-mover
findface-multi-findface-multi-file-mover-1
Started
After the installation is complete, the following output is shown in the console:
Tip
Be sure to save this data: you will need it later.
############################################################################# # Installation is complete # ############################################################################# - all configuration and data is stored in /opt/findface-multi - upload your license to http://192.168.0.90/#/license/ - user interface: http://192.168.0.90/ superuser: admin documentation: http://192.168.0.90/doc/ - superset BI: http://192.168.0.90:8088/ 2023/08/01 08:00:22 Installer finished
Perform the post-deployment procedures.
Tip
To install the same configuration of FindFace Multi on a different host, use the automatic deployment from the installation file. In this case, you won’t have to answer the installation questions again. The exact path to a JSON file with answers is displayed right after the last question, before the installer starts active progress:
[I 2023-08-01 07:56:36,323 main:142] Your answers were saved to /tmp/findface-installer-t8qk_isw.json
Be sure to remove fields *.config, exp_ip, and int_ip before installing FindFace Multi on a host with a different IP address.
To launch the automatic deployment from the /tmp/<findface-installer-*>.json file, execute:
sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.json
FindFace Video Worker Deployment on Remote Hosts
You can install findface-video-worker instances apart from the FindFace Multi principal server when creating a multi-host environment.
Important
Before deploying findface-video-worker instances on remote hosts, do the following:
Allow accessing the
findface-ntlslicense server from any IP address. To do so, open the/opt/findface-multi/configs/findface-ntls/findface-ntls.yamlconfiguration file on the server withfindface-ntlsand setlisten: 0.0.0.0:3133. Restart thefindface-multi-findface-ntls-1container.sudo vi /opt/findface-multi/configs/findface-ntls/findface-ntls.yaml listen: 0.0.0.0:3133
sudo docker container restart findface-multi-findface-ntls-1
Allow accessing the
findface-video-managerservice from any IP address. To do so, open the/opt/findface-multi/configs/findface-video-manager/findface-video-manager.yamlconfiguration file on the server withfindface-video-managerand setlisten: 0.0.0.0:18810andrpc:listen: 0.0.0.0:18811. Restart thefindface-multi-findface-video-manager-1container.sudo vi /opt/findface-multi/configs/findface-video-manager/findface-video-manager.yaml listen: 0.0.0.0:18810 ... rpc: listen: 0.0.0.0:18811
sudo docker container restart findface-multi-findface-video-manager-1
On the FindFace Multi server, open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file and make sure that theROUTER_URLparameter contains the external IP address of the FindFace Multi server and not the localhost. Thefindface-video-workerinstances on the remote hosts will be using this address for posting objects.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py ... 'ROUTER_URL': 'http://192.168.0.12', ...
To install only a findface-video-worker service, do the following:
Tip
Before deployment, be sure to meet the system requirements and prepare the server first.
Download the installer file
findface-*.run.Put the
.runfile into some directory on the designated host (for example,/home/username).From this directory, make the
.runfile executable.Note
Be sure to specify the actual file name instead of
findface-*.chmod +x findface-*.run
Execute the
.runfile.sudo ./findface-*.run
The installer will ask you a few questions and perform several automated checks to ensure that the host meets the system requirements. After filling out each prompt, press Enter. The questions and answers are the following:
Q:
Which product should be installed?A:
31. [multi ] FindFace Multi 2. [server ] FindFace Server 3. [video-worker] FindFace Video Worker (default: multi) product> 3
Q:
Which variant of Video Worker should be installed?A: Specify the
findface-video-workerpackage type, CPU or GPU.Which variant of Video Worker should be installed? - 1 [cpu] CPU-based implementation, slower but doesn't require GPU - 2 [gpu] CUDA-based implementation of video detector, requires NVIDIA GPU
Q:
Found X interface(s). Which one should we announce as our inter-service communication address?A: Choose the default interface
2to connect to the FindFace Multi server. We don’t recommend you to use localhost.Found 1 interface(s). Which one should we announce as our external address? - 1 [lo ] 127.0.0.1 - 2 [ens3 ] 192.168.112.254 (default: 192.168.112.254) inter_ip.advertised> 2
Q:
Please enter FF.Multi or FF.Server IP address:A: Specify the FindFace Multi server IP address.
Please enter FF.Multi or FF.Server IP address: server_addr> 192.168.112.25
After that, the installation process will automatically begin.
Perform post-installation procedures.
Configure the
findface-video-worker.yamlfile and fill out the values for all required parameters. Make sure thatdetectors,normalizers, andextractorsare specified in themodelssection andobjectssection has corresponding values. Below is an example of how the section should look like. It may vary depending on the recognition objects that you have selected.sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: cache_dir: /var/cache/findface/models_cache detectors: face: fnk_path: /usr/share/findface-data/models/detector/face.jasmine_fast.003.cpu.fnk min_size: 60 body: fnk_path: /usr/share/findface-data/models/detector/body.jasmine_fast.018.cpu.fnk min_size: 60 car: fnk_path: /usr/share/findface-data/models/detector/car.jasmine_fast.005.cpu.fnk min_size: 60 normalizers: face_norm: fnk_path: /usr/share/findface-data/models/facenorm/crop2x.v2_maxsize400.cpu.fnk face_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/crop1x.v2_maxsize400.cpu.fnk body_norm: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk body_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk car_norm: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk car_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk extractors: face_quality: fnk_path: /usr/share/findface-data/models/faceattr/quality_fast.v1.cpu.fnk normalizer: face_norm_quality body_quality: fnk_path: /usr/share/findface-data/models/pedattr/pedattr.quality.v0.cpu.fnk normalizer: body_norm_quality car_quality: fnk_path: /usr/share/findface-data/models/carattr/carattr.quality.v0.cpu.fnk normalizer: car_norm_quality objects: face: normalizer: face_norm quality: face_quality track_features: '' body: normalizer: body_norm quality: body_quality track_features: '' car: normalizer: car_norm quality: car_quality track_features: ''
Restart the
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
Important
If you chose to install findface-ntls and/or findface-video-manager on different hosts than that with findface-multi-legacy, specify their IP addresses in the /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml configuration file after the installation.
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml
In the ntls-addr parameter, specify the findface-ntls host IP address.
ntls_addr: 127.0.0.1:3133
In the mgr-static parameter, specify the findface-video-manager host IP address, which provides findface-video-worker with settings and the video stream list.
mgr:
static: 127.0.0.1:18811
Tip
To install a findface-video-worker instance on a different host, use the automatic deployment from the installation file. In this case, you won’t have to answer the installation questions again. The exact path to the installation file is displayed right after the last question before the installer starts active progress:
[I 2023-02-09 11:13:37,187 main:142] Your answers were saved to /tmp/findface-installer-p01n9sn3.json
To launch the automatic deployment from the /tmp/<findface-installer-*>.json file, execute:
sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.json
If your findface-video-worker instance is deployed in a multi-GPU environment, you will have to manually configure the /opt/findface-multi/docker-compose.yaml file.
In the
/opt/findface-multi/docker-compose.yamlfile in theservicessection, specify thefindface-video-workerconfiguration for each runningfindface-video-workerinstance. In theenvironmentparameter, specify the correct GPU ID forCUDA_VISIBLE_DEVICES. Stick to the example below:sudo vi /opt/findface-multi/docker-compose.yaml services: ... findface-video-worker: command: [--config=/etc/findface-video-worker.yaml] environment: [CUDA_VISIBLE_DEVICES=0] image: docker.int.ntl/ntech/universe/video-worker-gpu:ffserver-9.230407.1 logging: {driver: journald} network_mode: host restart: always runtime: nvidia volumes: ['./configs/findface-video-worker/findface-video-worker.yaml:/etc/findface-video-worker.yaml:ro', './models:/usr/share/findface-data/models:ro', './cache/findface-video-worker/models:/var/cache/findface/models_cache', './cache/findface-video-worker/recorder:/var/cache/findface/video-worker-recorder'] findface-video-worker-1: command: [--config=/etc/findface-video-worker-1.yaml] environment: [CUDA_VISIBLE_DEVICES=1] image: docker.int.ntl/ntech/universe/video-worker-gpu:ffserver-9.230407.1 logging: {driver: journald} network_mode: host restart: always runtime: nvidia volumes: ['./configs/findface-video-worker/findface-video-worker-1.yaml:/etc/findface-video-worker-1.yaml:ro', './models:/usr/share/findface-data/models:ro', './cache/findface-video-worker/models:/var/cache/findface/models_cache', './cache/findface-video-worker/recorder:/var/cache/findface/video-worker-recorder']
In the same section, specify the environment variables for each
findface-video-workerinstance. Stick to the example below:findface-video-worker: image: docker.int.ntl/ntech/universe/video-worker-gpu:ffserver-9.230407.1 ... environment: - CFG_MGR_STATIC=vm:18811 - CFG_NTLS_ADDR=ntls:3133 - CFG_STREAMER_PORT=9999 - CFG_STREAMER_URL=127.0.0.1:9999 - CFG_STREAMER_TRACKS=true #- CFG_RESOLUTIONS=1920x1080
Important
For the correct operation, multiple
findface-video-workerinstances should not refer to the sameCFG_STREAMER_PORT. Make sure that you have specified a uniqueCFG_STREAMER_PORTfor eachfindface-video-workerinstance.From the
/opt/findface-multidirectory rebuild all FindFace Multi containers.cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Deploy Video Recorder Step-by-Step
There are the following ways to deploy Video Recorder:
Automatically during the FindFace Multi deployment from a console installer. See FindFace Multi Standalone Automated Deployment for details.
Step-by-step. Recommended in a multi-host environment.
This section will guide you through the Video Recorder step-by-step deployment. Follow the instructions below minding the sequence.
Tip
Be sure to learn the FindFace Multi architecture first.
In this section:
Install APT Repository
First of all, install the FindFace apt repository as follows:
Download the installer file
findface-*.run.Put the
.runfile into some directory on the designated host (for example,/home/username).From this directory, make the
.runfile executable.chmod +x findface-*.run
Execute the
.runfile.sudo ./findface-*.run
The installer will ask you a few questions and perform several automated checks to ensure that the host meets the system requirements. After filling out each prompt, press Enter. The questions and answers are the following:
Q:
Which product should be installed?A:
1Which product should be installed? 1. [multi ] FindFace Multi 2. [server ] FindFace Server 3. [video-worker] FindFace Video Worker (default: multi) product> 1Q:
Please choose installation type:A:
3- 1 [stand-alone ] Single Server - 2 [multi-worker] Single Server, Multiple video workers - 3 [images ] Don't configure or start anything, just load the images and copy the models - 4 [custom ] Fully customized installation (default: stand-alone) type> 3Q:
Directory to install into:A: Specify the FindFace Multi installation directory. By default, it’s
/opt/findface-multi. Press Enter to confirm it. Otherwise, specify the directory of your choice and press Enter.Directory to install into: (default: /opt/findface-multi) dest_dir>Q:
Select models to install. Note that you will need to accordingly edit extraction-api and tntapi configuration files. At least one of recognition models has to be enabled.A: By default, all neural network models are subject to installation. You can leave it as is by entering
done, or select specific models. To do so, deselect all those on the list by entering-*in the command line, then select the required models by entering their sequence numbers (keyword): for example,1 3 4. Enterdoneto save your selection and proceed to another step. If FindFace Multi is deployed earlier, the models do not need to be installed.Select models to install. Note that you will need to accordingly edit extraction-api and tntapi configuration files. At least one of recognition models has to be enabled. - 1 [ ] ./models/carattr/carattr.categories.v0.cpu.fnk ... - 91 [ ] ./models/pedrec/pedrec.clio.gpu.fnk Enter keyword to select matching choices or -keyword to clear selection. Enter "done" to save your selection and proceed to another step. findface-data.models> doneAfter that, the FindFace apt repository will be automatically installed.
Install Main Components
To install the Video Recorder components, do the following:
Open the
/opt/findface-multi/docker-compose.yamlconfiguration file. Add thefindface-video-storageandfindface-video-streamerservices.sudo vi /opt/findface-multi/docker-compose.yaml ... findface-video-storage: command: [--config=/etc/findface-video-storage.conf] depends_on: [mongodb] image: docker.int.ntl/ntech/universe/video-storage:ffserver-9.230407.1 network_mode: service:pause restart: always volumes: ['./configs/findface-video-storage/findface-video-storage.yaml:/etc/findface-video-storage.conf:ro'] findface-video-streamer: command: [--config=/etc/findface-video-streamer-cpu.ini] depends_on: [findface-ntls, mongodb] image: docker.int.ntl/ntech/universe/video-streamer-cpu:ffserver-9.230407.1 network_mode: service:pause restart: always volumes: ['./configs/findface-video-streamer/findface-video-streamer.yaml:/etc/findface-video-streamer-cpu.ini:ro', './cache/findface-video-streamer:/var/cache/findface/video-streamer']
Create directories.
sudo mkdir -p /opt/findface-multi/configs/findface-video-storage/ /opt/findface-multi/configs/findface-video-streamer/
Create a configuration file
/opt/findface-multi/configs/findface-video-storage/findface-video-storage.yamland add the following:sudo vi /opt/findface-multi/configs/findface-video-storage/findface-video-storage.yaml listen: :18611 debug: false external-address: http://ip_address:18611/ streamer: endpoints: - 127.0.0.1:9000 chunk-storage: type: webdav webdav: timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 10s max-idle-conns-per-host: 20 keepalive: 24h0m0s trace: false upload-url: http://127.0.0.1:3333/uploads/video_storage s3: endpoint: '' bucket-name: '' access-key: '' secret-access-key: '' secure: true region: '' public-url: '' operation-timeout: 30 localfs: directory: '' meta-storage: mongo-uri: mongodb://127.0.0.1 database: video-storage timings: connect: 3s
Note
<ip_address>in the sectionexternal-addressmust be substituted into your ip address for FindFace Multi.Create a configuration file
/opt/findface-multi/configs/findface-video-streamer/findface-video-streamer.yamland add the following:sudo vi /opt/findface-multi/configs/findface-video-streamer/findface-video-streamer.yaml streamer: port: 9000 max_backpressure: 33554432 io_buffer_size: 524288 video_storage: url: http://127.0.0.1:18611 timeout: 6 cache: dir: /var/cache/findface/video-streamer
Start FindFace Multi containers:
cd /opt/findface-multi/ docker-compose up -d
Video Recorder is now successfully deployed. To configure Video Recorder after the deployment, follow these instructions.
Fully Customized Installation
The FindFace Multi developer-friendly installer provides you with a few installation options, including the fully customized installation. This option is mostly used when deploying FindFace Multi in a highly distributed environment and requires a high level of knowledge and experience.
To initiate the fully customized installation, do the following:
Download the installer file
findface-*.run.Put the
.runfile into some directory on the designated host (for example,/home/username).From this directory, make the
.runfile executable.Note
Be sure to specify the actual file name instead of
findface-*.chmod +x findface-*.run
Execute the
.runfile.sudo ./findface-*.run
The installer will ask you a few questions and perform several automated checks to ensure that the host meets the system requirements. After filling out each prompt, press Enter. The questions and answers are the following:
Q:
Which product should be installed?A:
1Which product should be installed? 1. [multi ] FindFace Multi 2. [server ] FindFace Server 3. [video-worker] FindFace Video Worker (default: multi) product> 1
Q:
Please choose installation type:A:
4Please choose installation type: - 1 [stand-alone ] Single Server - 2 [multi-worker] Single Server, Multiple video workers - 3 [images ] Don't configure or start anything, just load the images and copy the models - 4 [custom ] Fully customized installation (default: stand-alone) type> 4
Q:
Directory to install into:A: Specify the FindFace Multi installation directory. The default suggestion is
/opt/findface-multi. Press Enter to confirm it. Otherwise, specify the directory of your choice and press Enter.Directory to install into: (default: /opt/findface-multi) dest_dir>
Q:
Please enter path to docker-compose binary:A: Specify the actual path to the
docker-composebinary. The default suggestion is/usr/local/bin/docker-composeand it is the path you get when you installdocker-composeas prescribed by this document. Press Enter to confirm it. Otherwise, specify another path and press Enter.Please enter path to docker-compose binary (default: /usr/local/bin/docker-compose) docker_compose>
Q:
Found X interface(s). Which one should we announce as our external address?A: Choose the interface that you are going to use as the instance IP address.
Found 1 interface(s). Which one should we announce as our external address? - 1 [lo ] 127.0.0.1 - 2 [ens3 ] 192.168.112.254 (default: 192.168.112.254) ext_ip.advertised> 2
Q:
Found X interface(s). Which one should we announce as our inter-service communication address?A: Choose the interface for inter-service communication.
Found 1 interface(s). Which one should we announce as our inter-service communication address? - 1 [lo ] 127.0.0.1 - 2 [ens3 ] 192.168.112.254 (default: 192.168.112.254) inter_ip.advertised> 2
Q:
Please select FindFace Multi components to install:A: Choose FindFace components to install. By default, all components are subject to installation. You can leave it as is by entering
done, or select specific components. Whenever you have to make a selection, first, deselect all the listed components by entering-*in the command line, then select required components by entering their sequence number (keyword), for example:1 7 13, etc. Enterdoneto save your selection and proceed to another step.Warning
We do not recommend excluding the
pausecomponent from installation, as this will leave the remaining components without a network namespace to attach to.If you exclude the
pausecomponent intentionally, make sure to edit the/opt/findface-multi/docker-compose.yamlfile and provide a host name for each service in thenetwork_modeparameter.Please select FindFace Multi components to install: - 1 [v] findface-data - Face recognition models ... ... Enter keyword to select matching choices or -keyword to clear selection. Enter "done" to save your selection and proceed to another step. components> done
Specific questions related to the selected components: acceleration type, the required number of component instances, neural network models, etc. If you are experiencing a difficulty answering them, try to find an answer in this documentation, or submit your question to support@ntechlab.com.
Q:
Please set findface-multi admin passwordA: Set the Super Administrator (
superuser) password.Please set findface-multi admin password findface-multi-admin-password> admin
The installer will pull the FindFace Multi images from the Ntechlab registry and start the associated services in Docker containers.
Perform the post-deployment procedures.
Tip
To install the same configuration of FindFace Multi on a different host, use the automatic deployment from the installation file. In this case, you won’t have to answer the installation questions again. The exact path to the installation file is displayed right after the last question, before the installer starts active progress:
[I 2023-02-09 11:13:37,187 main:142] Your answers were saved to /tmp/findface-installer-p01n9sn3.json
Be sure to remove fields *.config, exp_ip, and int_ip before installing FindFace Multi on a host with a different IP address.
To launch the automatic deployment from the /tmp/<findface-installer-*>.json file, execute:
sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.json
Guide to Typical Multi-Host Deployment
This section is all about deploying FindFace Multi in a multi-host environment.
Tip
If after having read this section, you still have questions, do not hesitate to contact our experts by support@ntechlab.com.
Important
This section doesn’t cover the Video Recorder deployment. You can find a step-by-step instruction on this subject here.
The reasons for deploying FindFace Multi in a multi-host environment are the following:
The necessity to distribute the video processing high load.
The necessity to process video streams from a group of cameras in the place of their physical location.
Note
The most common use cases where such need comes to the fore are hotel chains, chain stores, several security checkpoints in the same building, etc.
The necessity to distribute the feature vector extraction high load.
Large number of objects to search through, that requires implementation of a distributed object database.
Before you start the deployment, outline your system architecture, depending on its load and allotted resources (see Requirements). The most common distributed scheme is as follows:
One principal server with the following components:
findface-ntls,findface-security,findface-sf-api,findface-video-manager,findface-upload,findface-video-worker,findface-extraction-api,findface-tarantool-server,pause, and third-parties.Several additional video processing servers with installed
findface-video-worker.(If needed) Several additional extraction servers with installed
findface-extraction-api.(If needed) Additional database servers with multiple Tarantool shards.
This section describes the most common distributed deployment. In high load systems, it may also be necessary to distribute the API processing (findface-sf-api and findface-video-manager) across several additional servers. This procedure requires a high level of expertise and some extra coding. Please do not hesitate to contact our experts for help (support@ntechlab.com).
To deploy FindFace Multi in a multi-host environment, follow the steps below:
Deploy Principal Server
To deploy the principal server as part of a distributed architecture, do the following:
On the designated physical server, install FindFace Multi from installer as follows (don’t forget to prepare a server prior the FindFace Multi deployment):
Product to install:
FindFace Multi.Installation type:
Single server, multiple video workers. In this case, FindFace Multi will be installed and configured to interact with additional remotefindface-video-workerinstances.Type of the
findface-video-workeracceleration (on the principal server): CPU or GPU, subject to your hardware configuration.Type of the
findface-extraction-apiacceleration (on the principal server): CPU or GPU, subject to your hardware configuration.
After the installation is complete, the following output will be shown on the console:
############################################################################# # Installation is complete # ############################################################################# - all configuration and data is stored in /opt/findface-multi - upload your license to http://172.20.77.17/#/license/ - user interface: http://172.20.77.17/ superuser: admin documentation: http://172.20.77.17/doc/
Upload the FindFace Multi license file via the main web interface
http://<Host_IP_address>/#/license. To access the web interface, use the providedsuperusercredentials.Note
The host IP address is shown in the links to FindFace web services in the following way: as an external IP address if the host belongs to a network, or
127.0.0.1otherwise.Important
Do not disclose the
superuser(Super Administrator) credentials to others. To administer the system, create a new user with the administrator privileges. Whatever the role, Super Administrator cannot be deprived of its rights.Allow the licensable services to access the
findface-ntlslicense server from any IP address, To do so, open the/opt/findface-multi/configs/findface-ntls/findface-ntls.yamlconfiguration file and setlisten: 0.0.0.0:3133. Restart thefindface-multi-findface-ntls-1container.sudo vi /opt/findface-multi/configs/findface-ntls/findface-ntls.yaml listen: 0.0.0.0:3133
sudo docker container restart findface-multi-findface-ntls-1
Allow accessing the
findface-video-managerservice from any IP address. To do so, open the/opt/findface-multi/configs/findface-video-manager/findface-video-manager.yamlconfiguration file and setlisten: 0.0.0.0:18810andrpc:listen: 0.0.0.0:18811. Restart thefindface-multi-findface-video-manager-1container.sudo vi /opt/findface-multi/configs/findface-video-manager/findface-video-manager.yaml listen: 0.0.0.0:18810 ... rpc: listen: 0.0.0.0:18811
sudo docker container restart findface-multi-findface-video-manager-1
Deploy Video Processing Servers
On an additional video processing server, install only a findface-video-worker instance following the step-by-step instructions. Answer the installer questions as follows:
Product to install: FindFace Video Worker.
Type of the
findface-video-workeracceleration: CPU or GPU, subject to your hardware configuration.FindFace Multi IP address: IP address of the principal server.
After that, the installation process will automatically begin. The answers will be saved to a file /tmp/<findface-installer-*>.json. Use this file to install FindFace Video Worker on other hosts without having to answer the questions again, by executing:
sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.jsonNote
If
findface-ntlsand/orfindface-video-managerare installed on a different host than that withfindface-security, specify their IP addresses in the/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file after the installation.sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlIn the
ntls-addrparameter, specify thefindface-ntlshost IP address.ntls-addr: 127.0.0.1:3133In the
mgr->staticparameter, specify thefindface-video-managerhost IP address, which providesfindface-video-workerwith settings and the video stream list.static: 127.0.0.1:18811
Deploy Extraction Servers
On an additional extraction server, install only a findface-extraction-api instance from the console installer. Answer the installer questions as follows:
Product to install:
FindFace Multi.Installation type:
Fully customized installation.FindFace Multi components to install:
findface-extraction-api,findface-data, andpause. To make a selection, first, deselect all the listed components by entering-*in the command line, then selectfindface-extraction-api,findface-data, andpauseby entering their sequence number (keyword). Enterdoneto save your selection and proceed to another step.Note
The
pausecomponent keeps information about other components’ network namespaces. It’s essential that you install it.Type of
findface-extraction-apiacceleration: CPU or GPU.Modification of the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file: specify the IP address of thefindface-ntlsserver.Neural network models to install: CPU or GPU model for face biometrics (mandatory), and (optional) CPU/GPU models to recognize face attributes, vehicle and vehicle attributes, and body and body attributes. Be sure to choose the right acceleration type for each model, matching the acceleration type of
findface-extraction-api: CPU or GPU. Be aware thatfindface-extraction-apion CPU can work only with CPU-models, whilefindface-extraction-apion GPU supports both CPU- and GPU-models.To move the principal
findface-extraction-apiinstance to another host, in the/opt/findface-multi/configs/findface-sf-api/findface-sf-api.yamlconfiguration file specify the IP address of the extraction server host and setlisten: 0.0.0.0:18411.listen: 0.0.0.0:18411 extraction-api: timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 10s max-idle-conns-per-host: 20 keepalive: 24h0m0s trace: false extraction-api: http://172.20.77.19:18666
After that, the installation process will automatically begin. The answers will be saved to a file /tmp/<findface-installer-*>.json. Use this file to install findface-extraction-api on other hosts without having to answer the questions again.
sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.json
After all the extraction servers are deployed, distribute load across them by using a load balancer.
Distribute Load across Extraction Servers
To distribute load across several extraction servers, you need to set up load balancing. The following step-by-step instructions demonstrate how to set up nginx load balancing in a round-robin fashion for 3 findface-extraction-api instances located on different physical hosts: one on the FindFace Multi principal server (172.168.1.9), and 2 on additional remote servers (172.168.1.10, 172.168.1.11). Should you have more extraction servers in your system, load-balance them by analogy.
Tip
You can use any load balancer according to your preference. Please refer to the relevant official documentation for guidance.
To set up load balancing, do the following:
Designate the FindFace Multi principal server (recommended) or any other server with running
findface-sf-apiservice as a gateway to all the extraction servers.Important
You will have to specify the gateway server IP address when configuring the FindFace Multi network.
On the designated server with the installed
findface-sf-apiinstance create anginxfolder that contains theextapi.conffile in the/opt/findface-multi/configs/directory. Make sure that theextapi.conffile includes the information like in the example below. In theupstreamdirective (upstream extapibackends), substitute the exemplary IP addresses with the actual IP addresses of the extraction servers. In theserverdirective, specify the gateway server listening port aslisten. You will have to enter this port when configuring the FindFace Multi network.upstream extapibackends { server 172.168.1.9:18666; ## ``findface-extraction-api`` on principal server server 172.168.1.10:18666; ## 1st additional extraction server server 127.168.1.11:18666; ## 2nd additional extraction server } server { listen 18667; server_name extapi; client_max_body_size 64m; location / { proxy_pass http://extapibackends; proxy_next_upstream error; } access_log /var/log/nginx/extapi.access_log; error_log /var/log/nginx/extapi.error_log; }
Define the Nginx service in the
docker-compose.yamlfile. To do that, add the container with Nginx image to thedocker-compose.yamlfile:sudo vi /opt/findface-multi/docker-compose.yaml nginx: image: nginx:latest ports: - 18667:18667 volumes: - ./configs/nginx/extapi.conf:/etc/nginx/conf.d/default.conf:ro
In the
findface-sf-apiconfiguration file, specify the distributor address:sudo vi /opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml listen: 0.0.0.0:18411 ... extraction-api: http://172.168.1.9:18667
Restart the containers.
cd /opt/findface-multi/ sudo docker-compose down sudo docker-compose up -d
On the principal server and each additional extraction server, open the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file. Substitute localhost in thelistenparameter with the relevant server address that you have specified inupstream extapibackends(/opt/findface-multi/configs/nginx/extapi.conf) before. In our example, the address of the 1st additional extraction server has to be substituted as such:sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml listen: 172.168.1.10:18666
Restart the
findface-multi-findface-extraction-api-1container on the principal server and each additional extraction server.sudo docker container restart findface-multi-findface-extraction-api-1
The load balancing is now successfully set up. Be sure to specify the actual gateway server IP address and listening port, when configuring the FindFace Multi network.
Deploy Additional Database Servers
The findface-tarantool-server component connects the Tarantool-based feature vector database and the findface-sf-api component, transferring search results from the database to findface-sf-api for further processing.
To increase search speed, you can allocate several additional servers to the feature vector database and create multiple findface-tarantool-server shards on each additional server. The concurrent functioning of multiple shards will lead to a remarkable increase in performance, as each shard can handle up to approximately 10,000,000 feature vectors.
To deploy additional database servers, do the following:
Install the
findface-tarantool-servercomponent on the first designated server. Thepausecomponent should already be installed on the server. If not, install it along with thefindface-tarantool-servercomponent. Answer the installer questions as follows:Product to install:
FindFace Multi.Installation type:
Fully customized installation.FindFace Multi components to install:
findface-tarantool-server,pause. To make a selection, first, deselect all the listed components by entering-*in the command line, then selectfindface-tarantool-serverandpauseby entering their sequence number (keyword). Enterdoneto save your selection and proceed to another step.
After that, the installation process will automatically begin.
As a result of the installation, the
findface-tarantool-servershards will be automatically installed in the amount ofN = min(max(min(mem_mb // 2000, cpu_cores), 1), 16 * cpu_cores). I.e., it is equal to the RAM size in MB divided by 2000, or the number of CPU physical cores (but at least one shard), or the number of CPU physical cores multiplied by 16 if the first obtained value is greater.Use the created
/tmp/<findface-installer-*>.jsonfile to installfindface-tarantool-serveron other servers without answering the questions again. To do so, execute:sudo ./<findface-*>.run -f /tmp/<findface-installer-*>.json
Be sure to specify the IP addresses and ports of the shards later on when configuring the FindFace Multi network. To learn the port numbers, execute on each database server:
sudo cat /opt/findface-multi/docker-compose.yaml | grep -E "CFG_LISTEN_PORT"
You will get the following result:
CFG_LISTEN_PORT=8101, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8102, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8103, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8104, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8105, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8106, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8107, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] CFG_LISTEN_PORT=8108, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()]
The port numbers are
8101,8102, etc.On the designated server with the installed
findface-tarantool-servercomponent, modify the configuration of each shard in the/opt/findface-multi/docker-compose.yamlfile. Specifyfindface-ntlslicense server address in theCFG_NTLSparameter. SetCFG_LISTEN_HOST=0.0.0.0.sudo vi /opt/findface-multi/docker-compose.yaml findface-tarantool-server-shard-001: depends_on: [] environment: ['TT_LISTEN=127.0.0.1:32001', TT_WORK_DIR=/var/lib/tarantool/FindFace, TT_WAL_DIR=xlogs, TT_MEMTX_DIR=snapshots, TT_MEMTX_MEMORY=2147483648, TT_CHECKPOINT_INTERVAL=14400, TT_CHECKPOINT_COUNT=3, TT_FORCE_RECOVERY=true, 'CFG_NTLS=172.23.218.110:3133', CFG_LISTEN_HOST=0.0.0.0, CFG_LISTEN_PORT=8101, CFG_EXTRA_LUA=loadfile("/tnt_schema.lua")()] image: docker.int.ntl/ntech/universe/tntapi:ffserver-9.230407.1 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./data/findface-tarantool-server/shard-001:/var/lib/tarantool/FindFace', './configs/findface-tarantool-server/tnt-schema.lua:/tnt_schema.lua:ro']
Restart the containers.
cd /opt/findface-multi/ sudo docker-compose down sudo docker-compose up -d
Open the
/opt/findface-multi/configs/findface-ntls/findface-ntls.yamlconfiguration file and setlisten: 0.0.0.0:3133. Restart thefindface-multi-findface-ntls-1container.sudo vi /opt/findface-multi/configs/findface-ntls/findface-ntls.yaml listen: 0.0.0.0:3133 license_dir: /ntech/license proxy: '' ui: 0.0.0.0:3185
sudo docker container restart findface-multi-findface-ntls-1
Modify the
/opt/findface-multi/configs/findface-sf-api/findface-sf-api.yamlconfiguration file. Setlisten: 0.0.0.0:18411and specify shards. Restart thefindface-multi-findface-sf-api-1container.sudo vi /opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml listen: 0.0.0.0:18411 extraction-api: timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 10s max-idle-conns-per-host: 20 keepalive: 24h0m0s trace: false extraction-api: http://127.0.0.1:18666 storage-api: timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 10s max-idle-conns-per-host: 20 keepalive: 24h0m0s trace: false shards: - master: http://172.20.77.19:8101/v2/ slave: '' - master: http://172.20.77.19:8102/v2/ slave: ''
sudo docker container restart findface-multi-findface-sf-api-1
To apply migrations, restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Configure Network
After all the FindFace Multi components are deployed, configure their interaction over the network. Do the following:
Open the
/opt/findface-multi/configs/findface-sf-api/findface-sf-api.yamlconfiguration file:sudo vi /opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml
Specify the following parameters:
Parameter
Description
extraction-api->extraction-apiIP address and listening port of the gateway extraction server with set up load balancing.
storage-api->shards->masterIP address and port of the
findface-tarantool-servermaster shard. Specify each shard by analogy.upload_urlWebDAV NginX path to send original images, thumbnails and normalized object images to the
findface-uploadservice.... extraction-api: extraction-api: http://172.168.1.9:18667 ... webdav: upload-url: http://127.0.0.1:3333/uploads/ ... storage-api: ... shards: - master: http://172.168.1.9:8101/v2/ slave: '' - master: http://172.168.1.9:8102/v2/ slave: '' - master: http://172.168.1.12:8101/v2/ slave: '' - master: http://172.168.1.12:8102/v2/ slave: '' - master: http://172.168.1.13:8102/v2/ slave: '' - master: http://172.168.1.13:8102/v2/ slave: ''
Restart the
findface-multi-findface-sf-api-1container.sudo docker container restart findface-multi-findface-sf-api-1
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Specify the following parameters:
Parameter
Description
SERVICE_EXTERNAL_ADDRESSFindFace Multi IP address or URL prioritized for the Genetec integration and webhooks. Once this parameter not specified, the system uses
EXTERNAL_ADDRESSfor these purposes. To use Genetec and webhooks, be sure to specify at least one of those parameters:SERVICE_EXTERNAL_ADDRESS,EXTERNAL_ADDRESS.EXTERNAL_ADDRESS(Optional) IP address or URL that can be used to access the FindFace Multi web interface. Once this parameter not specified, the system auto-detects it as the external IP address. To access FindFace Multi, you can use both the auto-detected and specified IP addresses.
VIDEO_DETECTOR_TOKENTo authorize the video object detection module, come up with a token and specify it here.
VIDEO_MANAGER_ADDRESSIP address of the
findface-video-managerhost.NTLS_HTTP_URLIP address of the
findface-ntlshost.ROUTER_URLExternal IP address of the
findface-securityhost that will receive detected objects from thefindface-video-workerinstance(s).SF_API_ADDRESSIP address of the
findface-sf-apihost.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py ... # SERVICE_EXTERNAL_ADDRESS is prioritized for FFSecurity webhooks and Genetec plugin. SERVICE_EXTERNAL_ADDRESS = 'http://localhost' EXTERNAL_ADDRESS = 'http://127.0.0.1' ... FFSECURITY = { 'VIDEO_DETECTOR_TOKEN': '7ce2679adfc4d74edcf508bea4d67208', ... 'VIDEO_MANAGER_ADDRESS': 'http://127.0.0.1:18810', ... 'NTLS_HTTP_URL': 'http://127.0.0.1:3185', 'ROUTER_URL': 'http://172.168.1.9', ... 'SF_API_ADDRESS': 'http://127.0.0.1:18411', ... }
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
The FindFace Multi components interaction is now set up.
Important
To preserve the FindFace Multi compatibility with the installation environment, we highly recommend you to disable the Ubuntu automatic update. In this case, you will be able to update your OS manually, fully controlling which packages to update.
To disable the Ubuntu automatic update, execute the following commands:
sudo apt-get remove unattended-upgrades
sudo systemctl stop apt-daily.timer
sudo systemctl disable apt-daily.timer
sudo systemctl disable apt-daily.service
sudo systemctl daemon-reload
Important
The FindFace Multi services log a large amount of data, which can eventually lead to disc overload. To prevent this from happening, we advise you to disable rsyslog due to its suboptimal log rotation scheme and use the appropriately configured systemd-journal service instead. See Logging for the step-by-step instructions.
Installation of Neural Network Models
To detect and recognize objects and object attributes, findface-extraction-api uses neural networks.
If you want to manually initiate the installation of neural network models, use the console installer as follows:
Execute the
findface-*file.Note
Be sure to specify the actual file name instead of
findface-*.sudo ./findface-*.run
Product to install:
FindFace MultiSelect the installation type:
Fully customized installation.Select a FindFace component to install:
findface-data. To do so, first, deselect all the listed components by entering-*in the command line, then select the required component by entering its sequence number (keyword). Enterdoneto save your selection and proceed to the next step.In the same manner, select models to install. After that, the installation process will automatically begin.
You can find installed models for the object and object attribute recognition at /opt/findface-multi/models/. See Neural Networks Summary.
Post-deployment Procedures and How-to’s
After you are finished with the FindFace Multi deployment, perform the procedures below.
In this section:
Specify Time Zone
The time zone on the FindFace Multi server determines the time in reports, logs, and names of such FindFace Multi artifacts as detection full frames and thumbnails, etc.
The time zone is specified in the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file, either in the Region/Country/City or Etc/GMT+H format. The best way to do so is to copy/paste your time zone from this table on Wikipedia.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
# time zone
TIME_ZONE = 'America/Argentina/Buenos_Aires'
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker container restart findface-multi-findface-multi-legacy-1
License Instance
FindFace Multi provides several licensing options. Whichever option you choose, you upload the FindFace Multi license file via the main web interface http://<Host_IP_address>/#/license. To access the web interface, use the superuser credentials.
Important
Do not disclose the superuser (Super Administrator) credentials to others. To administer the system, create a new user with administrator privileges. Whatever the role, the Super Administrator cannot be deprived of its rights.
Refer to the Licensing section to learn about the licensing options available.
Configure Logging
By default, the FindFace Multi processes are logged to Docker container logs, which can be accessed via the docker logs and docker service logs commands. In addition, Docker uses the json-file logging driver, which caches container logs in JSON. You can configure Docker to use another logging driver, choosing from the multiple logging mechanisms available. See Logging to learn how to do it.
Useful Docker Commands
In order to efficiently and easily administer FindFace Multi, you must have extensive knowledge and skills with Docker. If you’re new to Docker, get started with the commands below. Then explore the Docker documentation for additional skills.
View all Docker containers, including the stopped ones:
docker ps -a
To get a more compact and understandable output, execute:
docker ps -a --format "table {{.ID}}\t{{.Names}}\t{{.Status}}\t{{.State}}"
To extend the previous output, execute:
docker ps --format='{{json .}}' | jq
Restart the Docker service:
sudo systemctl restart docker
View a container log if the
journaldlogging driver is enabled:journalctl CONTAINER_NAME=findface-multi-findface-multi-legacy-1 -f
Stop a Docker container:
sudo docker container stop <container_name>/<container_id>
Stop all Docker containers:
sudo docker container stop $(sudo docker ps -a -q)
Start a Docker container:
sudo docker container start <container_name>/<container_id>
Start all Docker containers:
sudo docker container start $(sudo docker ps -a -q)
The FindFace Multi
docker-compose.yamlfile can be viewed as such:cat /opt/findface-multi/docker-compose.yaml
FindFace Multi configuration files can be found here:
cd /opt/findface-multi/configs/
Once you made changes to a configuration file, restart a relevant container by executing:
sudo docker container restart <container_name>/<container_id>
Enter a running Docker container to execute a command in it:
sudo docker container exec -it <container_name> /bin/bash
Stop and remove all FindFace Multi containers:
cd /opt/findface-multi sudo docker-compose down
Build, recreate, and start FindFace Multi containers:
cd /opt/findface-multi sudo docker-compose up -d
Remove FindFace Multi Instance
If your FindFace Multi instance is deployed as a standalone server in one run, use the uninstall.sh script integrated into the installer file to remove your instance.
Important
Make sure to back up your instance before uninstalling it if you plan to restore FindFace Multi and its data later on.
Run the script:
cd /opt/findface-multi/
sudo /opt/findface-multi/uninstall.sh /opt/findface-multi/
You will be asked a question:
Q: This script will remove all docker containers, images, volumes, config files and directory "/opt/findface-multi/?". Do you want to proceed(y/n)?:
Answer y to remove the instance.
In case your FindFace Multi deployment is not typical, and you are unable to use the automatic uninstaller script, remove all product components manually. The following command replicates the uninstall.sh script behavior and should be executed on every employed host:
sudo docker-compose -f /opt/findface-multi/docker-compose.yaml down -v --rmi all
sudo rm -rf /opt/findface-multi
Administration and Basic Configuration
Licensing
In this chapter:
Licensing Principles
The FindFace Multi licensing is granted using the following criteria:
The overall number of extracted feature vectors, regardless of the object type (face, body, vehicle).
Note
The feature vectors are extracted from objects detected in the video, from images in the record index, and user photos, and when building so-called cluster centroids.
The licensing scheme is the following:
Events: 1 event of video object detection = 1 object in a license.
Record Index: 1 photo in a record = 1 object in a license.
Clusters: 1 person = 1 object in a license.
Users: 1 photo of a user = 1 object in a license.
The number of video sources currently in use (i.e., active video processing jobs for cameras and video files).
The number of model instances in use in the
findface-extraction-apiservice.Face attribute recognition: gender/age/emotions/glasses/beard/face mask/etc.
Body attribute recognition: clothing color/type/etc.
Vehicle attribute recognition: make/model/color/body style/etc.
License plate recognition.
Face liveness detection.
Video recording.
Integrations with partners.
Integration with external VMS.
You can choose between the following licensing methods:
The online licensing is provided by interaction with the NtechLab Global License Manager
license.ntechlab.comand requires a stable internet connection, DNS, and open port 443 TCP. Upon being disconnected from the internet, the system will continue working off-grid for 4 hours.
Note
It is possible to prolongate the off-grid period for up to 2 days. Inform your manager if you need that.
The offline licensing via a USB dongle requires a USB port on the physical server with the
findface-ntlsservice (license server in the FindFace core).The offline licensing via hardware fingerprint requires Sentinel drivers installed on the physical server with the
findface-ntlsservice.
Important
For the system to function, a single instance of findface-ntls should be enough. If your system requires more license servers, contact your NtechLab manager beforehand to prevent your system from being blocked.
View and Update License
After installing FindFace Multi, upload the license file you obtained from the manager into the system. To do so, navigate to .
Use the same tab to consult current licensing information and upgrade your license.
Offline Licensing via USB dongle
To implement the licensing via a USB dongle, do the following:
Inform your manager that you intend to apply this licensing method and request your USB dongle and a license file.
Open the
/opt/findface-multi/docker-compose.yamlconfiguration file.sudo vi /opt/findface-multi/docker-compose.yaml
Add the line
privileged: true. Mount the/devdirectory into thefindface-multi-findface-ntls-1container by listing it in the volumes of thefindface-ntlssection. As a result, the entire section will look as follows:findface-ntls: command: [--config=/etc/findface-ntls.cfg] image: docker.int.ntl/ntech/universe/ntls:ffserver-8.221216 network_mode: service:pause privileged: true restart: always user: root volumes: ['./configs/findface-ntls/findface-ntls.yaml:/etc/findface-ntls.cfg:ro', './data/findface-ntls:/ntech/license', '/dev:/dev']
Create a new udev rule.
Download the
95-grdnt.rulesfile (e.g., into the/home/username/tmp/directory).Copy the
95-grdnt.rulesfile into the/etc/udev/rules.d/directory.sudo cp /home/username/tmp/95-grdnt.rules /etc/udev/rules.d/
Rebuild all FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Insert the USB dongle into a USB port.
Upload the license file on the License tab.
Offline Licensing via Hardware Fingerprint
Note
Sentinel is a type of offline licenses that, unlike guardant licenses, do not require any physical media for its work.
Glossary:
Sentinel is a software protection and licensing system by Thales. It allows you to implement offline licensing without access to a global server.
The C2V file is a file, containing data about a hardware fingerprint of the client’s machine, for binding the license only to this machine. This file is generated by the sentinel library. The C2V file is generated on the client’s machine where the license key will be installed later.
To implement the fingerprint licensing, do the following:
Inform your manager that you intend to apply this licensing method and request your unique license id. The manager will also supply you with the
findface-sentinel-lib_*.debpackage necessary for the FindFace Multi and Sentinel integration.Install the Sentinel drivers on the physical server with the
findface-ntlscomponent.Do the following:
Download Sentinel drivers from the official website.
Unzip the downloaded archive and browse to it.
tar -xvzf Sentinel_LDK_Linux_Runtime_Installer_script.tar.gz cd Sentinel_LDK_Linux_Runtime_Installer_script/
There is another archive
aksusbd-8.31.1.tar.gzinside the archive. Unzip it and browse to the resulting directory.tar -xvzf aksusbd-8.31.1.tar.gz cd aksusbd-8.31.1/
Run the installation command.
sudo ./dinst
Run and check the statuses of the Sentinel services.
sudo systemctl start aksusbd.service hasplmd.service sudo systemctl status aksusbd.service hasplmd.service
Mount the
/var/hasplmand/etc/hasplmdirectories into thefindface-multi-findface-ntls-1container. To do so, open the/opt/findface-multi/docker-compose.yamlconfiguration file and list them in the volumes of thefindface-ntlssection.sudo vi /opt/findface-multi/docker-compose.yaml findface-ntls: ... volumes: ['./configs/findface-ntls/findface-ntls.yaml:/etc/findface-ntls.cfg:ro', './data/findface-ntls:/ntech/license', '/var/hasplm:/var/hasplm', '/etc/hasplm:/etc/hasplm']
Rebuild all FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Put the
findface-sentinel-lib_*.debpackage received from your manager into some directory on the same host. Install the package.sudo dpkg -i /path/to/findface-sentinel-lib_*.deb
In the FindFace Multi web interface, navigate to . Take a hardware fingerprint (C2V file) by clicking the Download C2V for activation button.
Tip
If you prefer working with the console, you can send the following API request to
findface-ntlsinstead:curl <findface-ntls-server-ip>:3185/c2v >my_pc.c2v
Send the License ID and the C2V file to your manager and receive your license file in return.
Upload the license file on the License tab.
General Settings
The FindFace Multi general settings determine your system functioning and resource consumption. Here they are:
generic confidence threshold
thumbnail JPEG quality
schedule for automatic events/episodes cleanup
The general settings for faces, bodies, and vehicles are provided separately, depending on the enabled recognition objects.
To configure the general settings, navigate Settings -> General. After you are finished with adjustments, click Update.
By default, VMS cleanup settings are disabled in the general settings. To work with VMS cleanup settings from the web interface, follow this instruction Enable VMS Cleanup Settings in the UI. As a result, VMS cleanup settings will appear in the general settings list.
Find the detailed explanation of each general setting below.
In this section:
Generic Confidence Threshold
FindFace Multi verifies that objects match (i.e. two faces belong to the same person), based on the pre-defined confidence threshold. The default threshold is set to the optimum value. If necessary, you can change it.
Note
The higher is the threshold, the less are chances that a wrong object will be positively verified, however, some valid photos may also fail verification.
Tip
You can configure the confidence thresholds individually for each camera group and watch list.
Important
The default generic confidence threshold is optimal for the majority of recognition cases. We do not recommend changing it on your own. Be sure to consult with our technical experts prior (support@ntechlab.com).
Thumbnail JPEG Quality
Subject to JPEG quality, thumbnails may take up a significant amount of disc volume. Use the General tab to configure the parameter.
Automatic Event and Episode Cleanup
Use the same tab to schedule automatically purging old events and related episodes from the database. For example, you can purge matched and unmatched events/episodes on different schedules and purge only full frames.
Important
If you want to disable events cleanup for a specific parameter from the list, set it to 0 (zero) Hours. The system will behave as if this parameter does not exist.
E.g., to avoid cleanup of matched events for an object, specify 0 (zero) in the Delete matched events older than parameter. Then, Update the settings.
Global VMS Cleanup Settings
After you have enabled VMS cleanup settings in the UI, you can configure the regular video cleanup directly from the web interface.
The following settings are available:
Clean up between tracks: remove those intervals between tracks that do not contain recognition events.
Save tracks of events: save tracks of events for a selected object (face, body, or vehicle).
Keep events before and after: a video interval in seconds before and after a track with events that is kept along with the track.
Cleanup threshold: a minimum interval between tracks in seconds. If an interval between tracks is below the threshold, the video is not removed. A minimum allowed value for this field is 360 seconds.
Clean up between tracks from: time period (in days) during which the entire video archive is kept. After this period, the intervals between tracks will be cleaned up as configured in the settings. The value can be greater than or equal to zero.
Clean up archive: enable the entire video archive cleanup.
Archive cleanup age: a number of days the video archive is kept in the system. After that period, the video archive is removed.
VMS cleanup settings in the general settings section apply globally to the system. To configure video cleanup for a single camera, switch to custom VMS cleanup settings.
User Management and System Security
Important
Although FindFace Multi provides tools to ensure its protection from unauthorized access, they are not replacing a properly configured firewall. Be sure to use a firewall to heighten the FindFace Multi network protection.
Role and User Management
In this chapter:
Predefined Roles
FindFace Multi provides the following predefined roles:
Administrator is granted full access to the FindFace Multi functionality, integrative and administrative tools.
Important
Whatever the role, the first administrator (Super Administrator) cannot be deprived of its rights.
Operator is granted full access to the FindFace Multi functionality.
User is granted rights to modify their profile and work with events and episodes. The other functions are available read-only.
You can change the predefined roles privileges, as well as create various custom roles.
Create Custom Role in UI
To create a custom role in the web interface, do the following:
Navigate Settings -> Roles.
Click + Add new role.
On the Information tab, specify the role name. Save the role.
After saving the role, you will see the following tabs appear next to the Information tab:
Watch Lists: role privileges for specific watch lists
Camera Groups: role privileges for specific camera groups
Permissions: role privileges for entire system functions and entities
Set role privileges, subject to your needs. Note that there is a distinction between role privileges for a specific watch list/camera group and a system entity with the name
watchlist/cameragroup. For example, if you setOfffor a certain camera group on the Camera Groups tab, users with this role won’t be able to work with this very group of cameras. Unchecking all checkboxes for thecameragroupentity on the Permissions tab will prevent users from viewing and working with all camera groups.The full list of the FindFace Multi entities which are used in the current version is as follows:
all_own_sessions: all sessions of the current user on different devicesNote
If relevant permissions for this entity are set, users will be able to view (
view) and close (delete) all their sessions on different devices. Otherwise, users will be only allowed to view and close their session on the current device. Working with sessions takes place on the Sessions tab (Settings).bodycluster: cluster of bodiesbodyevent: body recognition eventbodyobject: full-length photo in a recordcamera: cameracameragroup: camera groupcarcard: vehicle recordcarcluster: cluster of vehiclescarepisode: vehicle-related episodecarevent: vehicle recognition eventcarobject: vehicle photo in a recordcounter: countersdeviceblacklistrecord: blocklistfacecluster: cluster of facesfaceevent: face recognition eventfaceobject: face photo in a recordhumancard: record of an individualhumanepisode: person-related episodereport: reportupload: item (photo) in batch photo uploaduploadlist: list of photos in batch uploaduser: uservideoarchive: object identification in video fileswatchlist: watch listwebhook: webhook
You can also enable and disable rights for the following functionality:
batchupload_cards: bulk record uploadchange_runtimesetting: changing the FindFace Multi general settingsview_auditlog: viewing and working with the audit logs.configure_ntls: configuration of thefindface-ntlslicense serverview_runtimesetting: viewing the FindFace Multi general settings
Save the changes.
Primary and Additional User Privileges
You assign privileges to a user by using roles:
Primary role: main user role, mandatory for assignment. You can assign only one primary role to a user.
An additional user role, optional for assignment. You can assign several roles to one user. The rights associated with the additional roles will be added to the primary privileges.
All users belonging to a particular primary role automatically get access to camera groups (and video archives within the group) and watch lists (and records in the watch list) created by a user with the same primary role, subject to the privileges defined by their additional role(s).
See also
Create User Account Manually
To create a user account manually, do the following:
Navigate Settings -> Users.
Click + Add new user.
On the Information tab, specify user data such as name, login, and password. If necessary, add a comment.
Note
When setting a password, mind password requirements:
at least 8 characters long
not only numerals
not within the list of 20000 commonly used passwords
not similar to other user attributes
only Latin letters, numerals, and special characters are allowed
From the Roles drop-down menu, select one or several user roles. Set one of them as the Primary role.
On the Photos tab, attach the user’s photo.
Save the user account.
Work with Roles and Users via Console
In case predefined roles have been removed from the system, use the following command to create them:
sudo docker exec -it findface-multi-findface-multi-identity-provider-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py create_groups
To create a user with Super Administrator rights (superuser), execute the following command, mind that password is a required argument:
sudo docker exec -it findface-multi-findface-multi-identity-provider-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py create_default_user --password <password>
Integrate with Active Directory for Auto User Creation
If there are many users in FindFace Multi, it can be inconvenient to create their accounts one by one. One of the ways to facilitate the user creation is to harness the FindFace Multi integration with Active Directory. To do so, follow the step-by-step instructions below, minding the sequence.
Log-in into the Active Directory server and do the following:
Create a new user account in the Active Directory domain to use as a service account.
Do the following:
Open Active Directory. Click Start, point to Administrative Tools, and then click Active Directory Users and Computers.
Click the domain name and then expand the contents. Right-click Users, point to New, and then click User. You will see a user creation form.
Fill-in the fields in the form at your discretion. On the second tab, check the Password never expires checkbox.
Click Next. Review the information that you provided, and if everything is correct, click Finish.
Right-click the just created user account, and then navigate Properties -> Member Of -> Add.
In the Select Groups dialog box, add the Domain Administrators and Domain Users groups to the list, and then click OK.
Click OK to finish.
Register a Service Principal Name (SPN) for the service account that you created. To do so, open PowerShell as administrator and execute the following command, specifying your actual
SERVICE USER NAMEand domain. In the example below, the domain name istestntl.local.setspn -A HTTP/<SERVICE USER NAME>.testntl.local@TESTNTL.LOCAL <SERVICE USER NAME>
In the same PowerShell window, generate a Keytab file by executing the command below with your actual
SERVICE USER NAME, domain, and desirableKEYTAB FILE NAME.ktpass.exe -princ HTTP/<SERVICE USER NAME>.testntl.local@TESTNTL.LOCAL -mapuser <SERVICE USER NAME> -crypto ALL -ptype KRB5_NT_PRINCIPAL -pass * -out c:\<KEYTAB FILE NAME>.keytab
To check the result, navigate to the root directory of the
Cdrive. You will see a keytab file with the relevant name.Move the keytab file that you created to the FindFace Multi server.
To successfully establish a link between FindFace Multi and Active Directory, you need to enable the Kerberos support in the findface-multi-findface-multi-ui-1 container on the FindFace Multi principal server. Do the following:
Mount the
/opt/findface-multi/configs/kerberos/krb5.conffile and the/opt/findface-multi/configs/keytabdirectory into thefindface-multi-findface-multi-ui-1container. To do so, open the/opt/findface-multi/docker-compose.yamlconfiguration file and list them in the volumes of thefindface-multi-uisection.sudo vi /opt/findface-multi/docker-compose.yaml findface-multi-ui: ... volumes: ['./configs/findface-multi-ui/nginx-site.conf:/etc/nginx/conf.d/default.conf:ro', './data/findface-multi-legacy/uploads:/var/lib/findface-security/uploads', './configs/kerberos/krb5.conf:/etc/krb5.conf:ro', './data/findface-multi-ui/keytab/:/keytab/']
Create directories for the mounted volumes:
/opt/findface-multi/configs/kerberos/and/opt/findface-multi/data/findface-multi-ui/keytab/. Copy the Kerberos configuration file from thefindface-multi-findface-multi-ui-1container to the/opt/findface-multi/configs/kerberos/directory.sudo mkdir -p /opt/findface-multi/data/findface-multi-ui/keytab/ sudo mkdir /opt/findface-multi/configs/kerberos/ sudo docker cp findface-multi-findface-multi-ui-1:/etc/krb5.conf /opt/findface-multi/configs/kerberos/
Open the
/opt/findface-multi/configs/kerberos/krb5.confconfiguration file. Specify the Active Directory realm in thelibdefaultssection. It must be equal to the Active Directory domain name, but spelled in upper case (TESTNTL.LOCALin the example below). Specify the Active Directory domain in therealmssection as well by analogy with the example below.sudo vi /opt/findface-multi/configs/kerberos/krb5.conf [libdefaults] default_realm = TESTNTL.LOCAL ... [realms] TESTNTL.LOCAL = { kdc = testntl.local default_domain = testntl.local }
Copy the keytab file to the
/opt/findface-multi/data/findface-multi-ui/keytab/directory.Append the following string to the
/etc/hostsfile:<Active Directory server IP address> <Active Directory domain name>.vi /etc/hosts ... 192.168.0.5 testntl.local
Open the
/opt/findface-multi/configs/findface-multi-ui/nginx-site.confconfiguration file. Find thelocation /users/me/adsection and uncomment it. Fill in the section by analogy with the example below, placing your actual variables in the strings with comments (#).The variables to specify are the following:
auth_gss_realm: realm name in Kerberosauth_gss_keytab: location of the keytab file.auth_gss_service_name: full service user name in Active Directory, including the name of the domain it belongs to
sudo vi /opt/findface-multi/configs/findface-multi-ui/nginx-site.conf location /users/me/ad { proxy_pass http://127.0.0.1/auth/ad_login/; # e.g http://127.0.0.1/auth/ad_login/; proxy_method POST; proxy_set_header X-Real-IP $remote_addr; proxy_set_header Host $http_host; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_set_header Authorization $http_authorization; proxy_pass_header Authorization; proxy_no_cache 1; proxy_cache_bypass 1; auth_gss on; auth_gss_realm TESTNTL.LOCAL; # e.g. TESTNTL.LOCAL; auth_gss_keytab /keytab/user.keytab; # e.g. /var/lib/web.keytab auth_gss_service_name HTTP/user.testntl.local; # e.g. HTTP/web.testntl.local; auth_gss_allow_basic_fallback on; }
To finalize the FindFace Multi integration with Active Directory, perform the following configuration steps on the FindFace Multi side:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
In the
SERVICESsection, set"active_directory": True.SERVICES = { ... "active_directory": True, ... } }
Fill in the
ACTIVE_DIRECTORY_CONFIGsection as follows:AUTH_LDAP_SERVER_URI:ldap: <Active Directory server IP address>AUTH_LDAP_BIND_DN: the name of the service user that you created in Active DirectoryAUTH_LDAP_BIND_PASSWORD: the service user passwordSEARCH_GROUPS: Active Directory organization units which FindFace Multi will search for user accounts
# Specify server credentials ACTIVE_DIRECTORY_CONFIG = { 'AUTH_LDAP_SERVER_URI': 'ldap://192.168.0.5', # Domain Administrator user 'AUTH_LDAP_BIND_DN': '<SERVICE USER NAME IN ACTIVE DIRECTORY>', # Domain Administrator user password 'AUTH_LDAP_BIND_PASSWORD': 'SERVICE USER NAME PASSWORD', # Specify organization units where users search will be executed. # Follow pattern (e.g. OU=DEV,DC=domain,DC=com) 'SEARCH_GROUPS': 'OU=DEV,DC=testntl,DC=local', }
Open the
/opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.pyconfiguration file and repeat the previous steps.Rebuild all FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
If the FindFace Multi integration with Active Directory is enabled, you will be able to set one of the Active Directory groups for a role you are creating or editing.
Once a user from the selected Active Directory group logs-in into FindFace Multi for the first time, they will be automatically added to the FindFace Multi user list.
To log-in with Active Directory, a user must click the Log in with Active Directory button in the authentication window, specify their Active Directory credentials, and click Sign in.
Deactivate or Delete Users
In order to deactivate a user, unset Active on the user list ().
If you are going to deactivate multiple users, select them on the user list and then click Deactivate selected.
To delete users from FindFace Multi, select them on the user list and then click Delete selected.
Authentication and Session Monitoring
In this section:
Authentication Types
FindFace Multi provides the following authentication types:
password: standard login/password authentication. Enabled by default.face: authentication is possible only by the user’s face.face_or_password: authentication is possible using either a face or login/password.face_and_password: two-factor authentication. After a face is successfully recognized, the user must enter their credentials.
Important
For all the authentication types based on face recognition, you need the following configuration:
standalone liveness service (
findface-liveness-api)
Important
Before using face recognition for authentication, you need to attach photos to users’ profiles and equip their workplaces with webcams.
Note
You can enable a work session monitoring for the authentication types face and face_or_password. In this case, the system will be periodically renewing the session after verifying that the face of a person at the workplace matches the user’s face that has logged in (see Configure Authentication and Session Renewal for details).
Tip
FindFace Multi also provides a certificate-based authentication that is configured independently. Contact our support team for details (support@ntechlab.com).
Configure Authentication and Session Renewal
To configure authentication and session monitoring, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. Find theFFSECURITYandFFSECURITY_AUTH_CONFIGsections.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FFSECURITY = { # auth config # available options: face, password, face_and_password, face_or_password 'AUTH_TYPE': 'face_or_password', # 180 days by default 'MAXIMUM_SESSION_LENGTH': 15552000, ... } ... # - FindFace Security authorization configuration dictionary - FFSECURITY_AUTH_CONFIG = { 'FACE_AUTH_CONFIDENCE': 0.740, # FAR = 2.5E-09 # model: [kiwi_320] # 3 settings below are for front-end only # session renew works only with face or face_or_password authorization type 'NEED_SESSION_RENEW': False, 'RENEW_SESSION_INTERVAL': 0, 'MAXIMUM_RENEW_ATTEMPTS': 2, }
In the
FFSECURITYsection, set the following authentication parameters:AUTH_TYPE: authentication type. Available options:face,password,face_and_password,face_or_password.MAXIMUM_SESSION_LENGTH: the maximum session length, in seconds. After a session expires, the user will be automatically logged out unless the session is renewed.
In the
FFSECURITY_AUTH_CONFIGsection, set the following authentication and session monitoring parameters:FACE_AUTH_CONFIDENCE: after a face in the webcam video is detected as alive, the system checks this face against the database of user photos with this confidence threshold.NEED_SESSION_RENEW: ifTrue, a session can be renewed and prolonged by the time equal toMAXIMUM_SESSION_LENGTH, after verifying that the face of a person at the workplace matches the user’s face that has logged in.RENEW_SESSION_INTERVAL: period in seconds before the expected time of the session expiry, during which the system will attempt to renew the session by enabling the webcam and verifying the user’s face.MAXIMUM_RENEW_ATTEMPTS: the number of user verification attempts. The attempts occur in a row during the renewal interval.
Note
A verification attempt takes about 3 seconds to complete.
Tip
We recommend you to set up the monitoring parameters so that
MAXIMUM_RENEW_ATTEMPTSmultiplied by the attempt duration is less thanRENEW_SESSION_INTERVAL. Otherwise, the system will extend the renewal interval x2, x3, and so on, subject to the number of attempts.Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Log out All Users
To automatically log out all users, execute the following command on the FindFace Multi principal server console:
sudo docker container exec -it findface-multi-findface-multi-identity-provider-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py logout_all_users
Tip
This command comes in handy when switching to a different authentication type.
Enable Data Encryption
To ensure data security, we recommend enabling SSL encryption. Do the following:
On the host system, create the nginx configuration directory with the subdirectory that will be used to store all the SSL data:
sudo mkdir -p /etc/nginx/ssl/
Create the SSL key and certificate files. When using self-signed certificate, use the following command:
sudo openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout /etc/nginx/ssl/my-example-domain.com.key -out /etc/nginx/ssl/my-example-domain.com.crt
You will be asked a few questions about your server in order to embed the information correctly in the certificate. Fill out the prompts appropriately. The most important line is the one that requests the
Common Name. You need to enter the domain name or public IP address that you want to be associated with your server. Both of the files you created (my-example-domain.com.keyandmy-example-domain.com.crt) will be placed in the/etc/nginx/ssl/directory.When using CA-certificate, add the certificate path to volumes for the
findface-video-workerservice, add the CA-certificates installation and update the root certificate store in the service container.Open the
docker-compose.yamlfile:sudo vi /opt/findface-multi/docker-compose.yaml
Locate the
findface-video-workersection and adjust it to make sure it looks as follows.For CPU:
findface-video-worker: entrypoint: ["sh", "-c", "apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends --yes ca-certificates && update-ca-certificates && exec /tini -- /findface-video-worker-cpu --config=/etc/findface-video-worker.yaml"] depends_on: [findface-video-manager, findface-ntls, mongodb] image: docker.int.ntl/ntech/universe/video-worker-cpu:ffserver-8.221216 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-video-worker/findface-video-worker.yaml:/etc/findface-video-worker.yaml:ro', './models:/usr/share/findface-data/models:ro', './cache/findface-video-worker/models:/var/cache/findface/models_cache', './cache/findface-video-worker/recorder:/var/cache/findface/video-worker-recorder', '/etc/nginx/ssl/my-example-domain.crt:/usr/local/share/ca-certificates/my-example-domain.crt:ro']
For GPU, it will be enough to add the path to the certificate and update the root certificate store:
findface-video-worker: entrypoint: ["sh", "-c", "update-ca-certificates && exec /tini -- /findface-video-worker-gpu --config=/etc/findface-video-worker.yaml"] depends_on: [findface-video-manager, findface-ntls, mongodb] environment: [CUDA_VISIBLE_DEVICES=0] image: docker.int.ntl/ntech/universe/video-worker-gpu:ffserver-8.221216 logging: {driver: journald} network_mode: service:pause restart: always runtime: nvidia volumes: ['./configs/findface-video-worker/findface-video-worker.yaml:/etc/findface-video-worker.yaml:ro', './models:/usr/share/findface-data/models:ro', './cache/findface-video-worker/models:/var/cache/findface/models_cache', './cache/findface-video-worker/recorder:/var/cache/findface/video-worker-recorder', '/etc/nginx/ssl/my-example-domain.crt:/usr/local/share/ca-certificates/my-example-domain.crt:ro']
Important
For CPU version, the configuration requires internet access. If there is no access, please, contact our support team (support@ntechlab.com).
Warning
For CPU version, the startup time will increase by ~15 seconds for the
findface-video-workercontainer.Rebuild all FindFace Multi containers.
cd /opt/findface-multi/ docker-compose down docker-compose up -d
Configure nginx to use SSL. Open the nginx configuration file
/opt/findface-multi/configs/findface-multi-ui/nginx-site.conf. Apply the following modifications to the file:Add the new
server {...}section that contains the URL replacement rule. In therewrite ^(.*) https://…line, replaceip_address_server_ffmultiwith IP address of the server where FindFace Multi is installed.server { listen 80; server_name my-example-domain.com www.my-example-domain.com; rewrite ^(.*) https://ip_address_server_ffmulti$1 permanent; access_log off; }Comment out the following lines in the existing
server {...}section:# listen 80 default_server; # listen [::]:80 default_server;
Add the following lines, including the paths to the certificate and the key, to the existing
server {...}section:listen 443 ssl; ssl_certificate /etc/nginx/ssl/my-example-domain.com.crt; ssl_certificate_key /etc/nginx/ssl/my-example-domain.com.key;
The example of the configuration file
/opt/findface-multi/configs/findface-multi-ui/nginx-site.confwith correctly configured SSL settings is shown below:upstream ffsecurity { server 127.0.0.1:8002; } upstream ffsecurity-ws { server 127.0.0.1:8003; } upstream ffsecurity-django { server 127.0.0.1:8004; } upstream audit { server 127.0.0.1:8012; } upstream identity-provider { server 127.0.0.1:8022; } map $http_upgrade $ffsec_upstream { default "http://ffsecurity-ws"; "" "http://ffsecurity"; } server { listen 80; server_name my-example-domain.com www.my-example-domain.com; rewrite ^(.*) https://my-example-domain.com$1 permanent; access_log off; } server { # listen 80 default_server; # listen [::]:80 default_server; listen 443 ssl; ssl_certificate /etc/nginx/ssl/my-example-domain.com.crt; ssl_certificate_key /etc/nginx/ssl/my-example-domain.com.key; root /var/lib/findface-security; autoindex off; server_name _; location = / { alias /usr/share/findface-security-ui/; try_files /index.html =404; } location /static/ { } location /uploads/ { # internal; # uncomment if you intend to enable OVERPROTECT_MEDIA add_header 'Access-Control-Allow-Origin' '*'; add_header 'Access-Control-Allow-Methods' '*'; add_header 'Access-Control-Allow-Headers' '*'; add_header 'Access-Control-Expose-Headers' 'Content-Length,Content-Range'; add_header 'Access-Control-Max-Age' 2592000; location ~ /card/(?<card_type>[a-zA-Z]+)/(?<card_id>[0-9]+)/attachments/(.*)$ { add_header 'Access-Control-Allow-Origin' '*'; add_header 'Access-Control-Allow-Methods' '*'; add_header 'Access-Control-Allow-Headers' '*'; add_header 'Access-Control-Expose-Headers' 'Content-Length,Content-Range'; add_header 'Access-Control-Max-Age' 2592000; add_header 'Content-Disposition' 'attachment'; add_header 'Content-Security-Policy' 'sandbox'; } } location /ui-static/ { alias /usr/share/findface-security-ui/ui-static/; } location /doc/ { alias /opt/findface-security/doc/; } location /api-docs { alias /opt/findface-security/rapidoc; index index.html; } location /api-docs/ { alias /opt/findface-security/rapidoc/; try_files $uri index.html =404; } location ~ /videos/(?<video_id>[0-9]+)/upload/(.*)$ { client_max_body_size 15g; proxy_set_header Host $http_host; proxy_set_header X-Forwarded-For $remote_addr; proxy_set_header X-Forwarded-Proto $scheme; proxy_pass http://ffsecurity; } location @django { internal; client_max_body_size 1g; proxy_set_header Host $http_host; proxy_set_header X-Forwarded-For $remote_addr; proxy_set_header X-Forwarded-Proto $scheme; proxy_http_version 1.1; proxy_read_timeout 5m; proxy_pass http://ffsecurity-django; } # location /v1/video-liveness { # add_header Access-Control-Allow-Headers "*" always; # add_header Access-Control-Allow-Methods "*" always; # add_header Access-Control-Allow-Origin "*" always; # # if ($request_method = 'OPTIONS') { # return 204; # } # # client_max_body_size 300m; # proxy_set_header Host $http_host; # proxy_set_header X-Forwarded-For $remote_addr; # proxy_set_header X-Forwarded-Proto $scheme; # proxy_pass http://127.0.0.1:18301; # proxy_read_timeout 5m; # } location / { client_max_body_size 1g; proxy_set_header Host $http_host; proxy_set_header X-Forwarded-For $remote_addr; proxy_set_header X-Forwarded-Proto $scheme; proxy_http_version 1.1; proxy_set_header Upgrade $http_upgrade; proxy_set_header Connection "upgrade"; proxy_pass $ffsec_upstream; proxy_read_timeout 5m; location ~ ^/(cameras|videos|vms|external-vms).*/stream/?$ { proxy_set_header Host $http_host; proxy_set_header X-Forwarded-For $remote_addr; proxy_set_header X-Forwarded-Proto $scheme; proxy_pass http://ffsecurity; } location ~ ^/streams/(.*)$ { internal; proxy_pass $1$is_args$args; } location /audit-logs { proxy_pass http://audit; } location ~ ^/(auth|ad_groups|cproauth|groups|permissions|sessions|users|user-face|device-blacklist-records) { proxy_pass http://identity-provider; } } # location /users/me/ad { # # proxy_pass <FFmulti_address>/auth/ad_login/; e.g http://127.0.0.1/auth/ad_login/; # proxy_method POST; # # proxy_set_header X-Real-IP $remote_addr; # proxy_set_header Host $http_host; # proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; # proxy_set_header Authorization $http_authorization; # proxy_pass_header Authorization; # proxy_no_cache 1; # proxy_cache_bypass 1; # # auth_gss on; # auth_gss_realm <REALM>; # e.g. TESTNTL.LOCAL; # auth_gss_keytab <path/to/file.keytab>; # e.g. /var/lib/web.keytab # auth_gss_service_name <service_name>; # e.g. HTTP/web.testntl.local; # auth_gss_allow_basic_fallback on; # } }Copy the generic nginx configuration file
nginx.conffrom thefindface-multi-findface-multi-ui-1container to the/etc/nginx/directory:sudo docker cp findface-multi-findface-multi-ui-1:/etc/nginx/nginx.conf /etc/nginx/nginx.conf
In the configuration file
/etc/nginx/nginx.conf, find theSSL Settingssection and append the following lines:ssl_session_cache shared:SSL:10m; ssl_session_timeout 1h;
In the
/opt/findface-multi/docker-compose.yamlfile, mount the SSL-encryption data directory/etc/nginx/ssl/and the configuration file/etc/nginx/nginx.confof the host system into thefindface-multi-findface-multi-ui-1container:Open the
docker-compose.yamlfile:sudo vi /opt/findface-multi/docker-compose.yaml
Locate the
findface-multi-uisection and adjust it to make sure it looks like this:findface-multi-ui: depends_on: [findface-multi-legacy] image: docker.int.ntl/ntech/multi/multi/ui:ffmulti-2.1.0 network_mode: service:pause restart: always volumes: ['./configs/findface-multi-ui/nginx-site.conf:/etc/nginx/conf.d/default.conf:ro', './data/findface-multi-legacy/uploads:/var/lib/findface-security/uploads', '/etc/nginx/ssl:/etc/nginx/ssl', '/etc/nginx/nginx.conf:/etc/nginx/nginx.conf:ro']
Edit the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.In the
ROUTER_URLandIMAGE_CROP_URLparameters, substitute thehttp://prefix withhttps://.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py ... 'ROUTER_URL': 'https://127.0.0.1', 'IMAGE_CROP_URL': 'https://127.0.0.1', ...
If you use a CA-certificate, specify in the
ROUTER_URLparameter the domain for which the certificate was created:'ROUTER_URL': 'https://my-example-domain.com'
Add
https://my-example-domain.comaddress to theEXTERNAL_ADDRESSparameter:... EXTERNAL_ADDRESS = 'https://my-example-domain.com' ...
Open the
/etc/hostsfile on the server where FindFace Multi is installed and add the following line:sudo vi /etc/hosts ... 127.0.0.1 my-example-domain.com
In the system where you use a browser to interact with FindFace Multi navigate to the
hostsfile. Add IP address of the server that hosts FindFace Multi instead of theip_address_server_ffmulti. Replacemy-example-domain.comwith your domain address - the same way you did it in the previous steps.For Linux OS do the following:
sudo vi /etc/hosts ... *ip_address_server_ffmulti* my-example-domain.com
If you use Windows OS, run
C:\Windows\System32\drivers\etc\hostsas an administrator. Add the following line to thehostsfile:*ip_address_server_ffmulti* my-example-domain.com
Restart the containers:
cd /opt/findface-multi/ sudo docker-compose down sudo docker-compose up -d
If you use self-signed certificate, disable SSL certificate verification for cameras and uploaded video archives:
Navigate to the Video Sources -> Cameras or Uploads.
Click to the camera or uploaded video archive.
On the Advanced tab, uncheck Verify the SSL certificate.
Warning
You may receive errors with getting screenshots from camera and working with video recorder, when deploying FindFace Multi in a highly distributed environment by fully customized installation and the pause component is excluded from installation. To resolve this problem, add the following lines to the end of the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py file:
USE_X_FORWARDED_HOST = True
SECURE_PROXY_SSL_HEADER = ('HTTP_X_FORWARDED_PROTO', 'https')
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker restart findface-multi-findface-multi-legacy-1
Enable Record Index Protection
If the record index security is turned off, photos and attachments in records will be available by direct link, no matter what the user’s rights are. Configure FindFace Multi to run all media requests through the DJANGO application for ACL checks to increase record index security.
Important
Enable the record media security only if you need it, as this setting severely negatively impacts the system performance.
Important
For the ACL checks to work properly, you must set the view permission for photos of faces, bodies, and vehicles stored in records. To do so, navigate Settings → Roles → specific role → Permissions and set View for the faceobject, bodyobject, and carobject entities, subject to the object types enabled in the system. See Create Custom Role in UI for details.
See also
To enable record index security, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. Locate theOVERPROTECT_MEDIAparameter and set itTrue.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py ... 'OVERPROTECT_MEDIA': True,
Do the same in the
/opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.pyconfiguration file: locate theOVERPROTECT_MEDIAparameter and set itTrue.sudo vi /opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.py ... 'OVERPROTECT_MEDIA': True,
Open the nginx configuration file
/opt/findface-multi/configs/findface-multi-ui/nginx-site.conf. Uncommentinternalin thelocation /uploadssection.sudo vi /opt/findface-multi/configs/findface-multi-ui/nginx-site.conf location /uploads/ { internal; # Uncomment if you intend to enable OVERPROTECT_MEDIA ... }
Restart the
findface-multi-findface-multi-legacy-1,findface-multi-findface-multi-identity-provider-1, andfindface-multi-findface-multi-ui-1containers.sudo docker container restart findface-multi-findface-multi-legacy-1 sudo docker container restart findface-multi-findface-multi-identity-provider-1 sudo docker container restart findface-multi-findface-multi-ui-1
After the new security policy is applied, logged-in users must re-authenticate. To make the users do so, execute the logout-all command:
sudo docker container exec -it findface-multi-findface-multi-identity-provider-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py logout_all_users
Disable ACL
You can turn off FindFace Multi ACL if you do not need it, as the constant permission checks consume a significant amount of system resources.
Do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Set
ENABLE_ACL = False.... ENABLE_ACL = False
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
List of User Sessions. Blocklist
In this chapter:
FindFace Multi allows you to monitor user sessions and learn associated data, such as the connected device UUID, type of user interface, IP address, last ping time, and so on.
If necessary, you can add a device to the blocklist without deactivating the user account. The device block may come in handy in various situations. For example, if you want users to access the system only from their workplaces. Use the blocklist functionality to take your system safety to the next level.
Grant Permissions to Work with Sessions
A user’s access to the list of sessions depends on the granted permissions:
Administrator: can view and close sessions of all users
User with the
all_own_sessionspermissions: can view/close all sessions initiated with their usernameUser without the
all_own_sessionspermissions: can only view/close their current session
View User Sessions
To view the list of user sessions, navigate Settings -> Sessions.
Each session record provides the following data:
device UUID
username
type of the user interface (mobile/web)
device information
IP address
status (online, offline, blocked)
last ping time
Use the filter panel above the list of sessions to set up the search conditions.
To close a session, select it in the list and click x.
Block Device
The list of blocked devices is available on the Blocklist Records tab.
You can add a device to the blocklist on the Sessions tab. Blocking a device leads to the user’s automatic log-out.
To block a device, do the following:
Select the relevant session record(s).
Click Block.
Specify the reason for the device to be blocked (mandatory) and the block expiry date (optional). If no date is specified, the block will be permanent.
Click Save.
Allowed File Extensions in Records
By default, you can attach a file of any extension to a record. It is possible to strengthen your system safety by creating the allowlist of file extensions. It will prevent your users from uploading files of unwanted formats, including those that might contain hidden malicious code, such as .js, .swf, and such.
To create the allowlist of file extensions, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
In the
FFSECURITYsection, find theCARD_ATTACHMENTS_FILENAME_REGEXPparameter. Set an expression with the allowed file extensions. Any valid Python regular expression will do.Examples:
r'.*\.png': allows only files with the.pngextensionr'.*\(png|jpg)': allows the.pngand.jpgextensionsr'.*': allows all file extensionsNone: allows all file extensions'XXXXXX': uploading files of any extension is prohibited
FFSECURITY = { ... 'CARD_ATTACHMENTS_FILENAME_REGEXP': r'.*\.txt', ... }
Tip
Commenting out the
CARD_ATTACHMENTS_FILENAME_REGEXPparameter also allows all file extensions.Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Camera Groups
Camera groups are entities that are used for video sources classification. After processing a video, the system will attribute the object recognition events obtained from the video to a designated camera group. It makes the further event handling and search a lot easier.
To create a camera group, do the following:
Navigate Settings -> Camera groups.
Click + Add new camera group.
On the Information tab, specify the group name. Add a comment if needed.
If you want to allocate a certain
findface-video-workerinstance to process video streams from the group, create or select one or several allocation labels.Note
To complete the allocation, list the labels in the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file. See Allocate findface-video-worker to Camera Group for details.If you want to deduplicate events from cameras that belong to the same group, i.e. exclude coinciding events, select Deduplicate events with interval and specify the deduplication interval in seconds (interval between 2 consecutive checks for event uniqueness).
Warning
Use deduplication with extreme caution. If cameras within a group observe different scenes, some objects may be skipped. See Deduplicate Events for details.
By default, video from all camera groups is processed using the generic confidence threshold. To set an individual threshold for the camera group, enable Confidence threshold and specify the threshold value.
Save the changes.
On the Permissions tab, assign privileges on the camera group, specifying which user roles are allowed to change/view the camera group settings.
Click Save.
Activate the camera group.
Allocate findface-video-worker to Camera Group
In a distributed architecture, it is often necessary that video streams from a group of cameras be processed in situ, without being redistributed across remote findface-video-worker instances by the principal server.
Note
Among typical use cases are hotel chains, chain stores, several security checkpoints in the same building, etc.
In this case, allocate the local findface-video-worker to the camera group.
Do the following:
Navigate to the Settings tab. Click Camera Groups.
Open the camera group settings.
In the Labels, create or select one or several allocation labels. Save changes.
There are two ways to specify allocation labels:
by
findface-video-worker.yamlconfiguration fileby
docker-compose.yamlconfiguration file
Note
Use the same name of labels in findface-video-worker and in the web-interface, otherwise cameras will not process.
Do the following:
Open the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file and specify the allocation labels in the dictionary format (labelsMyLabel1,MyLabel2in the example below).sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml labels: {"MyLabel1": "true", "MyLabel2": "true"}
Restart the
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
As an alternative way, you can use the configuration of docker-compose.yaml file:
Open the
/opt/findface-multi/docker-compose.yamlconfiguration file and specify the allocation labels in thefindface-video-workersection by adding environment:sudo vi /opt/findface-multi/docker-compose.yaml ... findface-video-worker: command: [--config=/etc/findface-video-worker.yaml] depends_on: [findface-video-manager, findface-ntls, mongodb] environment: - CUDA_VISIBLE_DEVICES=0 - CFG_LABLES=MyLabel1=true;MyLabel2=true image: docker.int.ntl/ntech/universe/video-worker-cpu:ffserver-9.230407.1 logging: {driver: journald} network_mode: service:pause ...
Rebuild all FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Note
If a camera is assigned an allocation label, its video stream can be processed by a findface-video-worker instance with the same label, as well as by all unlabeled findface-video-worker instances.
Warning
If a labeled camera is processed by an unlabeled findface-video-worker instance and a free similar-labeled instance appears, the camera won’t automatically switch to the latter. Restart the similar-labeled camera.
Watch Lists
The appearance of specific individuals and vehicles in the video is monitored with a set of default and custom watch lists.
Records of individuals and vehicles are assigned to watch lists. Once a watch list is activated, the system will be looking for each person or a vehicle on it during video processing.
You can create as many custom watch lists as necessary: wanted, suspects, etc. — subject to your needs.
In this section:
Monitoring Unmatched Faces
FindFace Multi features a special pre-configured watch list used for monitoring only unmatched events. This watch list cannot be removed from the system. To edit its settings, navigate to the Settings tab. Click Watch Lists and then click Unmatched.
Create Watch List
To create a custom watch list, do the following:
Navigate Settings -> Watch Lists.
Click + Add new watch list.
On the Information tab, specify the watch list name.
From the Color palette, select a color which will be shown in notifications for this list.
Describe the watch list in a comment if needed.
By default, all watch lists in the system are applied the generic confidence threshold. To set an individual threshold for the watch list, check Confidence threshold and specify the threshold value.
Important
The default generic confidence threshold is optimal for the majority of recognition cases. We do not recommend changing it on your own. Be sure to consult with our technical experts prior (support@ntechlab.com).
Enable Require event acknowledgment if it is mandatory that events associated with the list be manually acknowledged.
Enable Do not create events to avoid creating events once an object is detected.
Enable Enable sound alert to turn on sound notifications for the list if needed.
On the Permissions tab, assign privileges on the watch list, specifying which user roles are allowed to change/view the watch list settings.
Activate and save the watch list.
Remove Watch List
To remove a custom watch list, delete records associated with it first. Otherwise, the system will return an error and will not delete the watch list.
Default Watch List and Unmatched watch list cannot be removed from the system.
Configure Saving Images in Reports
When building reports, you will be able to choose to save the report images as links, thumbnails, or full frames. It is possible to configure the image parameters. To do so, open the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file and alter the default JPEG quality and the maximum height of thumbnails and full frames, subject to your free disc space.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
# reports image saving options
'REPORT_THUMBNAIL_JPEG_QUALITY': 75,
'REPORT_THUMBNAIL_MAX_HEIGHT': 100,
'REPORT_FULLFRAME_JPEG_QUALITY': 75,
'REPORT_FULLFRAME_MAX_HEIGHT': 250,
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker container restart findface-multi-findface-multi-legacy-1
Deduplicate Events
In this section:
Consider enabling deduplication to exclude coinciding object recognition events within one camera group.
Enable Deduplication
To enable event deduplication, do the following:
Enable the offline video detection mode for each camera in the group. See Add Camera for details.
Navigate to the Settings tab. Click Camera Groups.
Open the camera group settings.
Select Deduplicate events with interval and specify the deduplication interval in seconds.
How It Works
The deduplication algorithm works as follows. In the offline mode, the server receives one best object snapshot per tracking session on a camera.
Note
A tracking session continues until an object disappears from the camera field.
If there are several tracking sessions on a camera(s) of a camera group within the specified deduplication interval, FindFace Multi handles the received snapshots in the following way:
If there is a match with a record within the preceding deduplication interval, FindFace Multi drops a newly acquired snapshot. Otherwise, it saves the snapshot to the database.
For unmatched objects, FindFace Multi considers both the similarity between objects and snapshot quality when performing deduplication. As a result, FindFace Multi drops all snapshots within the deduplication interval unless a new object snapshot is of higher quality. Thus, it guarantees the system deduplicates events without skipping high-quality objects, which are essential for further video analytics.
Enable Personal Data Protection
FindFace Multi supports laws related to the processing of personal data of individuals (GDPR and similar).
To apply personal data protection to your system, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Disable saving unmatched events by setting
'IGNORE_UNMATCHED': True.... FFSECURITY = { ... # do not save unmatched events (GDPR support) 'IGNORE_UNMATCHED': True, ... }
For events with matches, enable blurring all unmatched objects in full frames. To do so, set
'BLUR_UNMATCHED_OBJECTS': True. Optionally, you can modify the default JPEG quality of those frames.... FFSECURITY = { ... # blur all unmatched objects on the full frame of the matched event (GDPR support) 'BLUR_UNMATCHED_OBJECTS': True, # full frame jpeg quality when `BLUR_UNMATCHED_OBJECTS` is enabled 'BLURRED_FULLFRAME_JPEG_QUALITY': 85, ... }
Enable blurring all unmatched objects on the Video Wall. To do so, set
"gdpr": Truein theFFSECURITY_UI_CONFIG->video_playersection.FFSECURITY_UI_CONFIG = { ... "video_player": { "overlay": { ... "gdpr": True } } ... }
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
See also
Configure Video Recorder
The FindFace Core internal architecture allows for embedding Video Recorder, an additional functionality that records, stores, and plays back video data from cameras.
This section is about the Video Recorder configuration.
In this section:
Enable Video Recorder
To enable Video Recorder, do the following:
Enable the
findface-video-workerservice to transfer video chunks to thefindface-video-storageservice.Important
This setting makes the Video Recorder functionality available over HTTP API. It’s also obligatory if you want to harness Video Recorder as part of the FindFace Multi web interface.
Do the following:
Open the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration to enablefindface-video-workerservice that will supply Video Recorder with video.sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml
Set
enabled: truein therecordersection.recorder: enabled: true ...
Restart
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
Configure Video Recorder to work as part of the FindFace Multi web interface.
Note
Omit the following steps if you do not need the Video Recorder tools to appear in the FindFace Multi web interface.
Do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Find the
FFSECURITY_UI_CONFIG->vmssection.The list of vms parameters to configure is the following:
Parameter
Description
vms->"enabled"Set
Trueto add the Record video checkbox to camera settings.vms->"video_player"Set
Trueto enable the video player camera settings to open in event notifications and camera preview (instead of static frames). This setting requires"enabled": True.Find the
FFSECURITY_UI_CONFIG->video_playersection.The list of video player parameters to configure is the following:
Parameter
Description
"overlay"->"objects"Defines the settings for displaying bbox and attribute data to the objects
"faces","bodies","cars". SetTrueorFalseto enable/disable bbox and attribute data display for the relevant objects."overlay"->"gdpr"Set
Trueto blurring all unmatched objects displayed on the video player. To fully comply with the personal data protection laws, follow this guide."timeline"->"min_zoom"Defines the largest possible scale on the video player timeline, seconds per pixel.
"timeline"->"max_zoom"Defines the smallest possible scale on the timeline, seconds per pixel.
"timeline"->"objects"->"events""faces"->"enabled"Set
Trueto enable marking face events on the timeline."timeline"->"objects"->"events""faces"->"limit"The maximum number of face events simultaneously marked on the timeline. If there is more face events than that, you will be asked to zoom in.
"timeline"->"objects"->"events""faces"->"matchedColor"Color of the matched face events.
"timeline"->"objects"->"events""faces"->"unmatchedColor"Color of the unmatched face events.
"timeline"->"objects"->"events""bodies"->"enabled"Set
Trueto enable marking bodies events on the timeline."timeline"->"objects"->"events""bodies"->"limit"The maximum number of body events simultaneously marked on the timeline. If there is more body events than that, you will be asked to zoom in.
"timeline"->"objects"->"events""bodies"->"matchedColor"Color of the matched body events.
"timeline"->"objects"->"events""bodies"->"unmatchedColor"Color of the unmatched body events.
"timeline"->"objects"->"events""cars"->"enabled"Set
Trueto enable marking vehicle events on the timeline."timeline"->"objects"->"events""cars"->"limit"The maximum number of vehicle events simultaneously marked on the timeline. If there is more vehicle events than that, you will be asked to zoom in.
"timeline"->"objects"->"events""cars"->"matchedColor"Color of the matched vehicle events.
"timeline"->"objects"->"events""cars"->"unmatchedColor"Color of the unmatched vehicle events.
"timeline"->"objects"->"episodes""humans"->"enabled"Set
Trueto enable marking human episodes on the timeline."timeline"->"objects"->"episodes""humans"->"limit"The maximum number of humans episodes simultaneously marked on the timeline. If there is more face episodes than that, you will be asked to zoom in.
"timeline"->"objects"->"episodes""humans"->"matchedColor"Color of the matched human episodes.
"timeline"->"objects"->"episodes""humans"->"unmatchedColor"Color of the unmatched human episodes.
"timeline"->"objects"->"episodes""cars"->"enabled"Set
Trueto enable marking vehicle episodes on the timeline."timeline"->"objects"->"episodes""cars"->"limit"The maximum number of vehicle episodes simultaneously marked on the timeline. If there is more vehicle episodes than that, you will be asked to zoom in.
"timeline"->"objects"->"episodes""cars"->"matchedColor"Color of the matched vehicle episodes.
"timeline"->"objects"->"episodes""cars"->"unmatchedColor"Color of the unmatched vehicle episodes.
FFSECURITY_UI_CONFIG = { "video_player": { "overlay": { "objects": { "faces": { "bbox": True, "info": True }, "bodies": { "bbox": True, "info": True }, "cars": { "bbox": True, "info": True } }, "gdpr": False }, "timeline": { "min_zoom": 0.2, "max_zoom": 200, "objects": { "events": { "faces": { "enabled": True, "limit": 500, "matchedColor": "rgba(6,193,103,0.8)", "unmatchedColor": "rgba(232,92,74,0.8)", }, "bodies": { "enabled": False, "limit": 500, "matchedColor": "rgba(6,193,103,0.8)", "unmatchedColor": "rgba(232,92,74,0.8)", }, "cars": { "enabled": False, "limit": 500, "matchedColor": "rgba(6,193,103,0.8)", "unmatchedColor": "rgba(232,92,74,0.8)", }, }, "episodes": { "humans": { "enabled": False, "limit": 500, "matchedColor": "rgba(6,193,103,0.8)", "unmatchedColor": "rgba(232,92,74,0.8)", }, "cars": { "enabled": False, "limit": 500, "matchedColor": "rgba(6,193,103,0.8)", "unmatchedColor": "rgba(232,92,74,0.8)", }, }, }, }, }, "vms": { "enabled": True, "video_player": True, },
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Nuances of Disabling Video Recorder
If Video Recorder is running and recording video on selected cameras, and you need to disable it, be sure to disable video recording on the cameras first, before proceeding with the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py and /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml configuration files.
Remove Video Records
See also
There are two methods of video archive cleanup available in the system:
Regular cleanup
Manual cleanup via a console command
In this section:
Regular Video Cleanup
To configure the regular video cleanup, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Find the
SERVICESsection and set"vms_cleanup": True.SERVICES = { "ffsecurity": { ... "vms_cleanup": True, } }
Find the
VMS_CLEANUP_SETTINGSsection and configure what video to remove, using the following parameters:'CLEANUP_BETWEEN_TRACKS': setTrueto remove video intervals that do not contain recognition events.'CLEANUP_ARCHIVE': setTrueto remove the entire video archive older than a given number of days.'ARCHIVE_CLEANUP_AGE': the maximum age of a video archive in the system, days. Applicable if'CLEANUP_ARCHIVE': True.'BETWEEN_TRACKS_CLEANUP_START_AGE': time period (days) during which the entire video archive is kept in the system. After this period is over, the system starts cleaning up between tracks according to theCLEANUP_BETWEEN_TRACKSsettings.'KEEP_EVENT_TYPES': event types that will be kept in the video archive, in the form of a dictionary, e.g., [‘face’, ‘body’, ‘car’].'KEEP_EVENT_BEFORE_AFTER': interval before and after a track with events that won’t be removed in order to preserve meaningful information related to them, seconds. The value must be less thanCLEANUP_THRESHOLD.'CLEANUP_THRESHOLD': minimum interval between tracks with events, seconds. If the interval between tracks is shorter, it won’t be removed.
'VMS_CLEANUP_SETTINGS': { 'CLEANUP_BETWEEN_TRACKS': True, 'CLEANUP_ARCHIVE': True, 'ARCHIVE_CLEANUP_AGE': 30, # days 'BETWEEN_TRACKS_CLEANUP_START_AGE': 0, # days 'KEEP_EVENT_TYPES': ['face', 'body', 'car'], # Add `safe_time_interval` in seconds to each `track` subject to delete # Prevents deletion of essential data. # Should be lower than CLEANUP_THRESHOLD 'KEEP_EVENT_BEFORE_AFTER': 10, # minimal allowed value # Threshold between `intervals` in seconds. # If duration between `tracks` < `CLEANUP_THRESHOLD` # interval between tracks will not be added to deletion tasks 'CLEANUP_THRESHOLD': 360, # minimal allowed value },
Find the
'VMS_CLEANUP_SERVICE_SCHEDULE'parameter and specify a recurrence rule (RRULE) for scheduling video cleanup.Tip
See the RRULE calculator here.
# rrule (recurrence rule) for scheduling `vms_cleanup` service 'VMS_CLEANUP_SERVICE_SCHEDULE': 'RRULE:FREQ=HOURLY;INTERVAL=3;WKST=MO',
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Manual Video Cleanup
You can manually remove video chunks older than the given number of days by executing the command below. To set the number of days, use the –vms-videos-max-age argument.
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup_vms --vms-videos-max-age=30
Enable VMS Cleanup Settings in the UI
By default, VMS cleanup settings are disabled in the UI. To work with VMS cleanup settings from the UI, configure the "vms" parameter in the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file.
In the
FFSECURITY_UI_CONFIGsection locate the"vms"parameter and set"enabled": True.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FFSECURITY_UI_CONFIG = { ... "vms": { "enabled": True, ... }, }
Make sure that VMS cleanup service is enabled in the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. Find theSERVICESsection and set"vms_cleanup": True.SERVICES = { "ffsecurity": { ... "vms_cleanup": True, } }
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Custom Tabs, Fields, and Filters in Record Index
See also
To create custom fields in the feature vector database, refer to Custom Metadata in Tarantool.
To add custom tabs and fields to the records of individuals and vehicles, do the following:
Prepare the list of custom tabs and fields you want to add to the records.
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Customize the records of individuals. To do so, uncomment the
FFSECURITY->CUSTOM_FIELDS->human_cardsection and modify the exemplary content, considering the following:'items': the list of fields in a record. Describe each field with the following parameters:'name': field’s internal name, string.'default': field’s default value. If a default value exceeds1e14 - 1, use a string data type to specify it, for example,"123123.."instead of123123...'label': field’s label in a record, string.'tab': tab that features the field.'display': display format (formorlist), string or array.'description': field’s description, string.'editable': field’s editability, boolean.'type': field data type, string. Possible values:list: requiresitems, additional parameter for lists (see below), expects objects {id, name} in dictionaries;valuelist: expects elements of primitive types.objectlist: allows for creating arrays of objects of required types.datetime: primitive data type displayed as a datetime list.date: primitive data type displayed as a date picker.boolean: primitive data type displayed as a checkbox.string: primitive data typestring.
additional parameters for lists (type=list, type=valuelist):
multiple: possibility of selecting several items in the list, boolean.items: dictionary used as a data source for the list.allow_create: possibility of adding new items to the list.custom_id: custom field for id (type=list).
additional parameters for object lists (type=objectlist).
object: objects used as a data source for the object list.simple: indicator that the field expects data of a primitive type instead of objects, for example, expects strings with phone numbers.
'filters': the list of search filters associated with the custom fields. Parameters:'name': filter’s internal name,'label': filter’s label in the web interface,'field': associated field in the format[field name].
'tabs': the list of tabs in a record.
FFSECURITY = { ... # -- Custom model fields -- # Edit CUSTOM_FIELDS -> `human_card` section to customize human card fields. # Edit CUSTOM_FIELDS -> `car_card` section to customize car card fields. ... 'CUSTOM_FIELDS': { 'human_card': { 'items': [ { 'name': 'personid', 'default': '', 'label': 'PersonID', 'display': ['list', 'form'], 'description': 'Sigur person ID', 'editable': False }, { 'name': 'firstname', 'default': '', 'label': 'First Name', 'display': ['list', 'form'], 'description': 'Sigur first name', 'editable': False }, { 'name': 'lastname', 'default': '', 'label': 'Last Name', 'display': ['list', 'form'], 'description': 'Sigur last name', 'editable': False }, { 'name': 'version', 'default': '', 'label': 'Version', 'display': ['list', 'form'], 'description': 'Sigur photo version', 'editable': False } ], 'filters': [ { 'name': 'personid', 'label': 'Sigur person ID filter', 'field': 'personid' } ] }, 'car_card': {}, # same fields are available }, }
Customize the vehicle records. To do so, duplicate the
human_cardsection content into thecar_cardsection and modify it by analogy.Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
You will see the custom content appear in the records.
Custom Metadata in Tarantool
It is often necessary to assign additional metadata to the objects extracted from images uploaded to the record index and now stored in the feature vector database.
In this section:
Customize Meta Fields of Face Objects
To assign custom meta fields to the face objects, do the following:
Prepare the list of custom meta fields to assign.
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
In the
FFSECURITYsection, uncomment theCUSTOM_FIELDS->face_objectsection and modify the exemplary content, considering the following:field_name: field’s name;type: data type (uint,stringorbool);default: field’s default value. If a default value exceeds ‘1e14 – 1’, use a string data type to specify it, for example,"123123.."instead of123123...
FFSECURITY = { ... # -- Custom model fields -- ... # Edit CUSTOM_FIELDS -> `face_object` section to customize face object fields. ... # 'CUSTOM_FIELDS': { ... 'face_object': { 'items': [ { "field_name": "tag_name_1", "type": "string", "default": "change_me" }, { "field_name": "tag_name_2", "type": "uint", "default": 123 }, { "field_name": "tag_name_3", "type": "bool", "default": True }, ] } }, }
Add the new meta fields to the feature vector database structure.
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
You can work with the new meta fields through HTTP API using the
objects/faces/methods.
Customize Meta Fields of Body and Vehicle Objects
Assigning custom meta fields to the body and vehicle objects are similar to the faces. The only difference is that you will need to manually create the CUSTOM_FIELDS -> body_object and CUSTOM_FIELDS -> car_object sections. It is a good idea to duplicate the CUSTOM_FIELDS -> face_object section and use it as a starting point for further modifications.
FFSECURITY = {
...
# -- Custom model fields --
...
# Edit CUSTOM_FIELDS -> `face_object` section to customize face object fields.
...
# 'CUSTOM_FIELDS': {
...
'body_object': {
'items': [
{
"field_name": "tag_name_1",
"type": "string",
"default": "change_me"
},
{
"field_name": "tag_name_2",
"type": "uint",
"default": 123
},
{
"field_name": "tag_name_3",
"type": "bool",
"default": True
},
]
}
'car_object': {
'items': [
{
"field_name": "tag_name_1",
"type": "string",
"default": "change_me"
},
{
"field_name": "tag_name_2",
"type": "uint",
"default": 123
},
{
"field_name": "tag_name_3",
"type": "bool",
"default": True
},
]
}
},
}
Similarly, you can work with the new meta fields using the objects/bodies/ and objects/cars/ methods of HTTP API.
Note
Assigning custom meta fields to the face, body and vehicle events are similar to the face, body and vehicle objects. You will also need to manually create the CUSTOM_FIELDS -> face_event, CUSTOM_FIELDS -> body_event and CUSTOM_FIELDS -> car_event sections by similar way.
See also
To create custom tabs, fields, and filters in cards, refer to Custom Tabs, Fields, and Filters in Record Index.
Console Bulk Record Upload
In addition to the web interface upload, you can bulk-upload records to the record index via the uploader.py console utility. We recommend preferring this utility over the web interface if the number of uploaded photos is more than 10,000.
Tip
To view the uploader.py help, execute:
docker exec findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py --help
Usage: uploader.py [OPTIONS] COMMAND [ARGS]...
Options:
--job FILE Job file (default: enroll-job.db)
--log-level TEXT Log level
--fsync BOOLEAN Call fsync() to prevent data loss on power failure
--help Show this message and exit.
Commands:
add Add items from CSV or TSV file to job
print Print contents of job file as JSON
run Run upload job
docker exec findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py add --help
Usage: uploader.py add [OPTIONS] FILES...
Options:
--format [csv|tsv] Input file format - CSV or TSV
--delimiter TEXT Field delimiter - by default it's "\t" for TSV and ","
for CSV
--help Show this message and exit.
docker exec findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py print --help
Usage: uploader.py print [OPTIONS]
Print contents of job file as JSON
Options:
--failed Show only failed images
--noface Show only images without detection
--help Show this message and exit.
docker exec findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py run --help
Usage: uploader.py run [OPTIONS]
Run upload job
Options:
--parallel INTEGER Number of enroll threads (default: 10)
--api TEXT API url (default: http://127.0.0.1:80/)
[required]
--user TEXT API username [required]
--password TEXT API password [required]
--watch-lists TEXT Comma-separated list of card list ids [required]
--inactive Mark new cards as inactive
--failed Include failed images
--noface Include images without detection
--all-faces Enroll all found faces on each image
--detect-timeout INTEGER Request timeout for detect photos
--logging-delta INTEGER Logging period delta
--help Show this message and exit.
Do the following:
Write the list of photos and metastrings to a CSV or TSV file.
Important
The file used as a metadata source must have the following format:
path to photo | metastring.To prepare a TSV file, you can use a
script, or generate it with python or other utility.Note
Both the script and the command in the examples below create the
images.tsvfile. Each image in the list will be associated with a metastring coinciding with the image file name in the formatpath to photo | metastring.To build a TSV file listing photos, upload the script to your home directory, for example
/home/ubuntu, and run the following command:sudo docker run -it --rm --network host --volume ${PWD}:/home/ubuntu/create_cards --volume /home/ubuntu/photos:/home/ubuntu/photos docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 sh -c "cd /home/ubuntu/create_cards && /opt/findface-security/bin/python3 tsv_builder.py /home/ubuntu/photos"where
/home/ubuntu/photosis a directory with your photos.A TSV file may also be generated using the following command:
python3 tsv_builder.py /home/ubuntu/photos/
You can use a
images.tsvfile as example.Create a job file out of a CSV or TSV file by using the
addutility command. As a result, a fileenroll-job.dbwill be created and saved in a current directory.sudo docker run -it --rm --network host --volume ${PWD}:/home/ubuntu/create_cards docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 sh -c "cd /home/ubuntu/create_cards && /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py add /home/ubuntu/create_cards/images.tsv"The
addutility options:--format: input file format, tsv by default,--delimiter: field delimiter, by default"\t"for TSV, and","for CSV.
Note
A job file represents a sqlite database which can be opened on the sqlite3 console.
Process the job file by specifying the path to the photos (for example,
/home/ubuntu/photos) and passing the necessary arguments. Use the following command:sudo docker run -it --rm --network host --volume ${PWD}:/home/ubuntu/create_cards --volume /home/ubuntu/photos:/home/ubuntu/photos docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 sh -c "cd /home/ubuntu/create_cards && /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py run --user admin --password password --watch-lists 1"The
runutility options:--parallel: the number of photo upload threads, 10 by default. The more threads you use, the faster the bulk upload is completed, however it requires more resources too.--api:findface-securityAPI URL,http://127.0.0.1:80/by default. Mandatory option.--user: login. Mandatory option.--password: password. Mandatory option.--watch-lists: comma-separated list of the watch lists id’s. Mandatory option.--inactive: mark new records as inactive.--failed: should an error occur during the job file processing, correct the mistake and try again with this option.--noface: by default, images classified as having no faces will be assigned theNOFACEstatus and automatically excluded from the upload. To attempt re-detecting faces in such images, re-run the job file with this option. If the re-detection gives a negative result again, an image will be skipped and a relevant record will appear in the upload log.--all-faces: upload all faces from a photo if it features several faces.--detect-timeout: request timeout for detect photos.--logging-delta: logging period delta.
(Optional) Print the job processing results as JSON. If necessary, you can print only failed images/ images without detected faces. Use the following command:
sudo docker run -it --rm --network host --volume ${PWD}:/home/ubuntu/create_cards --volume /home/ubuntu/NTL_impersonated:/home/ubuntu/NTL_impersonated docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 sh -c "cd /home/ubuntu/create_cards && /opt/findface-security/bin/python3 /tigre_prototype/ffsecurity/uploader.py print --noface --failed"The
printutility options:--failed: show only failed images.--noface: show only images without detection.
Configure FindFace Multi Neural Networks
Neural Networks Summary
Here you can find a summary of neural network models created by our Lab and used in FindFace Multi.
You can find installed models at /opt/findface-multi/models/.
Important
The default face biometrics model upon a clean install is mango_320.
Face, vehicle, and body detection
ls /opt/findface-multi/models/detector/
body.gustav_accurate.019.cpu.fnk body.jasmine_fast_nomany.018.gpu.fnk car.gustav_normal.004.gpu.fnk
body.gustav_accurate.019.gpu.fnk car.gustav_accurate.006.cpu.fnk face.jasmine_fast.003.cpu.fnk
body.jasmine_fast.018.cpu.fnk car.gustav_accurate.006.gpu.fnk face.jasmine_fast.003.gpu.fnk
body.jasmine_fast.018.gpu.fnk car.gustav_normal.004.cpu.fnk headbodyface.alpha000_normal.001.gpu.fnk
Face and body image normalization
ls /opt/findface-multi/models/facenorm/
bee.v3.cpu.fnk bee_fast.gpu.fnk crop2x.v2_maxsize400.cpu.fnk crop2x.v2_no_maxsize.gpu.fnk facenorm.multicrop_full_center_size400.cpu.fnk facenorm.multicrop_full_crop2x_size400.gpu.fnk
bee.v3.gpu.fnk crop1x.v2_maxsize400.cpu.fnk crop2x.v2_maxsize400.gpu.fnk cropbbox.v2.cpu.fnk facenorm.multicrop_full_center_size400.gpu.fnk
bee_fast.cpu.fnk crop1x.v2_maxsize400.gpu.fnk crop2x.v2_no_maxsize.cpu.fnk cropbbox.v2.gpu.fnk facenorm.multicrop_full_crop2x_size400.cpu.fnk
Face recognition
ls /opt/findface-multi/models/face/
mango_320.cpu.fnk mango_320.gpu.fnk nectarine_m_160.cpu.fnk nectarine_m_160.gpu.fnk nectarine_xs_320.cpu.fnk nectarine_xs_320.gpu.fnk
Face attribute recognition
ls /opt/findface-multi/models/faceattr/
age.v2.cpu.fnk beard.v0.gpu.fnk gender.v2.cpu.fnk glasses3.v0.gpu.fnk liveness.goodwin.cpu.fnk liveness.pacs.v2.gpu.fnk medmask3.v2.cpu.fnk quality_fast.v1.gpu.fnk
age.v2.gpu.fnk emotions.v1.cpu.fnk gender.v2.gpu.fnk headpose.v3.cpu.fnk liveness.goodwin.gpu.fnk liveness.web.v0.cpu.fnk medmask3.v2.gpu.fnk
beard.v0.cpu.fnk emotions.v1.gpu.fnk glasses3.v0.cpu.fnk headpose.v3.gpu.fnk liveness.pacs.v2.cpu.fnk liveness.web.v0.gpu.fnk quality_fast.v1.cpu.fnk
Vehicle image normalization
ls /opt/findface-multi/models/carnorm/
briacon.v0.cpu.fnk briacon.v0.gpu.fnk
Vehicle recognition
ls /opt/findface-multi/models/carrec/
bottas.cpu.fnk bottas.gpu.fnk
Vehicle attribute recognition
ls /opt/findface-multi/models/carattr/
carattr.categories.v0.cpu.fnk carattr.license_plate.v7.gpu.fnk carattr.orientation.v0.cpu.fnk carattr.quality.v0.gpu.fnk carattr.weight_types7.v0.cpu.fnk description.v0.gpu.fnk
carattr.categories.v0.gpu.fnk carattr.license_plate_quality.v1.cpu.fnk carattr.orientation.v0.gpu.fnk carattr.special_types11.v1.cpu.fnk carattr.weight_types7.v0.gpu.fnk
carattr.license_plate.v7.cpu.fnk carattr.license_plate_quality.v1.gpu.fnk carattr.quality.v0.cpu.fnk carattr.special_types11.v1.gpu.fnk description.v0.cpu.fnk
Body recognition
ls /opt/findface-multi/models/pedrec/
pedrec.durga.cpu.fnk pedrec.durga.gpu.fnk
Body attribute recognition
ls /opt/findface-multi/models/pedattr/
pedattr.age_gender.v0.cpu.fnk pedattr.bags.v0.cpu.fnk pedattr.clothes_type.v0.cpu.fnk pedattr.color.v1.cpu.fnk pedattr.protective.v1.cpu.fnk pedattr.quality.v0.cpu.fnk
pedattr.age_gender.v0.gpu.fnk pedattr.bags.v0.gpu.fnk pedattr.clothes_type.v0.gpu.fnk pedattr.color.v1.gpu.fnk pedattr.protective.v1.gpu.fnk pedattr.quality.v0.gpu.fnk
Enable Face and Face Attribute Recognition
FindFace Multi allows you to recognize human faces and face attributes. Subject to your needs, you can enable recognition of such face attributes as age, gender, emotions, beard, glasses, medical masks, head position, or liveness.
Face and face attribute recognition can be automatically enabled and configured during the FindFace Multi installation. If you skipped this step, you can manually do it later. Face and face attribute recognition works on both GPU- and CPU-acceleration.
Face object recognition is enabled by default. In case you removed face as a recognition object during the installation, you can add it later by following the below steps. If face object recognition is already installed, and you only need to enable face attribute recognition, jump to the steps 1.5, 1.6 and 4.1, 4.2. Other steps should be skipped.
To enable face recognition, do the following:
Specify neural network models for face object detection in the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file.Important
Be sure to choose the right acceleration type for each model, matching the acceleration type of
findface-extraction-api: CPU or GPU. Be aware thatfindface-extraction-apion CPU can work only with CPU-models, whilefindface-extraction-apion GPU supports both CPU- and GPU-models.Open the
findface-extraction-api.yamlconfiguration file.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
Specify the face detector model in the
detectors -> modelssection by pasting the following code:GPU
detectors: ... models: ... face_jasmine: aliases: - face - nnd - cheetah model: detector/face.jasmine_fast.003.gpu.fnk options: min_object_size: 32 resolutions: - 256x256 - 384x384 - 512x512 - 768x768 - 1024x1024 - 1536x1536 - 2048x2048 ...
CPU
detectors: ... models: ... face_jasmine: aliases: - face - nnd - cheetah model: detector/face.jasmine_fast.003.cpu.fnk options: min_object_size: 32 resolutions: - 256x256 - 384x384 - 512x512 - 768x768 - 1024x1024 - 1536x1536 - 2048x2048 ...
Make sure that the
objects -> facesection contains thequality_attribute: face_qualityand thebase_normalizer: facenorm/crop2x.v2_maxsize400.gpu.fnkor thebase_normalizer: facenorm/crop2x.v2_maxsize400.cpu.fnk, depending on your acceleration type:GPU
objects: ... face: base_normalizer: facenorm/crop2x.v2_maxsize400.gpu.fnk quality_attribute: face_quality ...
CPU
objects: ... face: base_normalizer: facenorm/crop2x.v2_maxsize400.cpu.fnk quality_attribute: face_quality ...
Specify the face normalizer models in the
normalizerssection by pasting the following code:GPU
normalizers: ... models: crop1x: model: facenorm/crop1x.v2_maxsize400.gpu.fnk crop2x: model: facenorm/crop2x.v2_maxsize400.gpu.fnk cropbbox: model: facenorm/cropbbox.v2.gpu.fnk multicrop_full_center: model: '' multicrop_full_crop2x: model: facenorm/facenorm.multicrop_full_crop2x_size400.gpu.fnk norm200: model: facenorm/bee.v3.gpu.fnk ...
CPU
normalizers: ... models: crop1x: model: facenorm/crop1x.v2_maxsize400.cpu.fnk crop2x: model: facenorm/crop2x.v2_maxsize400.cpu.fnk cropbbox: model: facenorm/cropbbox.v2.cpu.fnk multicrop_full_center: model: '' multicrop_full_crop2x: model: facenorm/facenorm.multicrop_full_crop2x_size400.cpu.fnk norm200: model: facenorm/bee.v3.cpu.fnk ...
Note
This step is required to enable face attribute recognition.
To enable face attribute recognition, do the following:
In the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file, specify the extraction models in theextractorssection, as shown in the example below. Be sure to indicate the right acceleration type for each model, matching the acceleration type offindface-extraction-api: CPU or GPU.GPU
extractors: ... models: face_age: faceattr/age.v2.gpu.fnk face_beard: faceattr/beard.v0.gpu.fnk face_beard4: '' face_countries47: '' face_emben: face/mango_320.gpu.fnk face_emotions: faceattr/emotions.v1.gpu.fnk face_eyes_attrs: '' face_eyes_openness: '' face_gender: faceattr/gender.v2.gpu.fnk face_glasses3: faceattr/glasses3.v0.gpu.fnk face_glasses4: '' face_hair: '' face_headpose: faceattr/headpose.v3.gpu.fnk face_headwear: '' face_highlight: '' face_liveness: faceattr/liveness.web.v0.gpu.fnk face_luminance_overexposure: '' face_luminance_underexposure: '' face_luminance_uniformity: '' face_medmask3: faceattr/medmask3.v2.gpu.fnk face_medmask4: '' face_mouth_attrs: '' face_quality: faceattr/quality_fast.v1.gpu.fnk face_scar: '' face_sharpness: '' face_tattoo: '' face_validity: ''
CPU
extractors: ... models: face_age: faceattr/age.v2.cpu.fnk face_beard: faceattr/beard.v0.cpu.fnk face_beard4: '' face_countries47: '' face_emben: face/mango_320.cpu.fnk face_emotions: faceattr/emotions.v1.cpu.fnk face_eyes_attrs: '' face_eyes_openness: '' face_gender: faceattr/gender.v2.cpu.fnk face_glasses3: faceattr/glasses3.v0.cpu.fnk face_glasses4: '' face_hair: '' face_headpose: faceattr/headpose.v3.cpu.fnk face_headwear: '' face_highlight: '' face_liveness: faceattr/liveness.web.v0.cpu.fnk face_luminance_overexposure: '' face_luminance_underexposure: '' face_luminance_uniformity: '' face_medmask3: faceattr/medmask3.v2.cpu.fnk face_medmask4: '' face_mouth_attrs: '' face_quality: faceattr/quality_fast.v1.cpu.fnk face_scar: '' face_sharpness: '' face_tattoo: '' face_validity: ''
The following extraction models are available:
Extractor
Acceleration
Configure as follows
age
CPU
face_age: faceattr/age.v2.cpu.fnkGPU
face_age: faceattr/age.v2.gpu.fnkbeard
CPU
face_beard: faceattr/beard.v0.cpu.fnkGPU
face_beard: faceattr/beard.v0.gpu.fnkindividual face feature vector
CPU
face_emben: face/mango_320.cpu.fnkGPU
face_emben: face/mango_320.gpu.fnkgender
CPU
face_gender: faceattr/gender.v2.cpu.fnkGPU
face_gender: faceattr/gender.v2.gpu.fnkemotions
CPU
face_emotions: faceattr/emotions.v1.cpu.fnkGPU
face_emotions: faceattr/emotions.v1.gpu.fnkglasses
CPU
face_glasses3: faceattr/glasses3.v0.cpu.fnkGPU
face_glasses3: faceattr/glasses3.v0.gpu.fnkhead position
CPU
face_headpose: faceattr/headpose.v3.cpu.fnkGPU
face_headpose: faceattr/headpose.v3.gpu.fnkface liveness
CPU
face_liveness: faceattr/liveness.web.v0.cpu.fnkGPU
face_liveness: faceattr/liveness.web.v0.gpu.fnkface mask
CPU
face_medmask3: faceattr/medmask3.v2.cpu.fnkGPU
face_medmask3: faceattr/medmask3.v2.gpu.fnkface quality
CPU
face_quality: faceattr/quality_fast.v1.cpu.fnkGPU
face_quality: faceattr/quality_fast.v1.gpu.fnkTip
To leave a model disabled, pass the empty value
''to the relevant parameter. Do not remove the parameter itself. Otherwise, the system will be searching for the default model.extractors: face_age: '' face_beard: '' face_beard4: '' face_countries47: '' face_emben: '' face_emotions: '' face_eyes_attrs: '' face_eyes_openness: '' face_gender: '' face_glasses3: '' face_glasses4: '' face_hair: '' face_headpose: '' face_headwear: '' face_highlight: '' face_liveness: '' face_luminance_overexposure: '' face_luminance_underexposure: '' face_luminance_uniformity: '' face_medmask3: '' face_medmask4: '' face_mouth_attrs: '' face_quality: '' face_scar: '' face_sharpness: '' face_tattoo: '' face_validity: ''
Important
The
face_livenessextraction modelliveness.web.v0is enabled by default. Do not disable it if you use authentication by face.Restart the
findface-multi-findface-extraction-api-1container.sudo docker container restart findface-multi-findface-extraction-api-1
To enable face recognition, modify the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file.In the
modelssection, specify the face neural network models by analogy with the example below:GPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: ... detectors: ... face: fnk_path: /usr/share/findface-data/models/detector/face.jasmine_fast.003.gpu.fnk min_size: 60 ... normalizers: ... face_norm: fnk_path: /usr/share/findface-data/models/facenorm/crop2x.v2_maxsize400.gpu.fnk face_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/crop1x.v2_maxsize400.gpu.fnk ... extractors: ... face_quality: fnk_path: /usr/share/findface-data/models/faceattr/quality_fast.v1.gpu.fnk normalizer: face_norm_quality
CPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: ... detectors: ... face: fnk_path: /usr/share/findface-data/models/detector/face.jasmine_fast.003.cpu.fnk min_size: 60 ... normalizers: ... face_norm: fnk_path: /usr/share/findface-data/models/facenorm/crop2x.v2_maxsize400.cpu.fnk face_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/crop1x.v2_maxsize400.cpu.fnk ... extractors: ... face_quality: fnk_path: /usr/share/findface-data/models/faceattr/quality_fast.v1.cpu.fnk normalizer: face_norm_quality
Make sure that the
objects -> facesection is included:objects: ... face: normalizer: face_norm quality: face_quality track_features: ''
Restart the
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
To enable face recognition, open the
/opt/findface-multi/configs/findface-video-manager/findface-video-manager.yamlconfiguration file and make sure it contains thefacesection indetectorsthat looks similar to the example below.sudo vi /opt/findface-multi/configs/findface-video-manager/findface-video-manager.yaml detectors: ... face: filter_min_quality: 0.5 filter_min_size: 60 filter_max_size: 8192 roi: '' fullframe_crop_rot: false fullframe_use_png: false jpeg_quality: 95 overall_only: true realtime_post_first_immediately: false realtime_post_interval: 1 realtime_post_every_interval: false track_interpolate_bboxes: true track_miss_interval: 1 track_overlap_threshold: 0.25 track_max_duration_frames: 0 track_send_history: false post_best_track_frame: true post_best_track_normalize: true post_first_track_frame: false post_last_track_frame: false tracker_type: simple_iou track_deep_sort_matching_threshold: 0.65 track_deep_sort_filter_unconfirmed_tracks: true track_object_is_principal: false track_history_active_track_miss_interval: 0
Note
This step is required to enable face attribute recognition.
Enable recognition of face attributes in the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.In the
FFSECURITYsection, specify the face attributes that you want to display for the face recognition events.# available features: age, beard, emotions, gender, glasses, headpose, medmask 'FACE_EVENTS_FEATURES': ['glasses', 'beard', 'age', 'gender', 'headpose', 'medmask', 'emotions'],
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Headpose Recognition
Headpose feature refers to the ability of the camera to detect and track the orientation and movement of a person’s head relative to the CCTV camera in real-time.
Warning
The headpose feature does not work when a person is wearing a medmask.
To detect the headpose (head turn and head tilt), FindFace Multi uses pitch and yaw.
Pitch refers to the rotation of the head around its horizontal axis, which runs from ear to ear. Positive pitch indicates that the head is tilted forward, while negative pitch indicates that the head is tilted backward.
Yaw refers to the rotation of the head around its vertical axis, which runs from top to bottom. Positive yaw indicates that the head is turned to the right, while negative yaw indicates that the head is turned to the left.
Pitch and yaw angles are measured relative to the camera and take on values from -90 to +90 degree.
You can see the value of head turn and head tilt in the attributes section of Events and filter face events by these parameters.
Possible scenarios of headpose recognition
Headpose recognition can be used in various scenarios where face recognition is used to improve accuracy and security, there are several of them:
Physical Access Control System, PACS: improving employee access control systems by ensuring that the face of the employee matches the expected orientation. It means that if a person is near the camera, turns the head to the camera, but is not going to pass through the PACS, the access will not be provided, due to the pre-set value of the head position.
Analytics of shopping center visitors: reducing the number of created low-quality face clusters. When a face cluster is forming, events corresponding to the specified values of the head rotation angles are taken into account. See more for clusters.
How to configure headpose
To configure headpose angle thresholds for creating clusters, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Enable the headpose threshold by setting
'FACE_CLUSTER_EVENT_HEADPOSE_THRESHOLDS_ENABLE': True,.If necessary, modify the lowest and highest threshold to pitch and yaw angles.
'FACE_CLUSTER_EVENT_YAW_ANGLE_LOWEST_THRESHOLD': -30, 'FACE_CLUSTER_EVENT_YAW_ANGLE_HIGHEST_THRESHOLD': 30, 'FACE_CLUSTER_EVENT_PITCH_ANGLE_LOWEST_THRESHOLD': -60, 'FACE_CLUSTER_EVENT_PITCH_ANGLE_HIGHEST_THRESHOLD': 60,
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Enable Vehicle and Vehicle Attribute Recognition
FindFace Multi allows you to recognize a single vehicle and its attributes.
The vehicle attributes are as follows:
license plate number (for selected countries),
Note
Currently, the following countries are supported:
Europe: Russia, Lithuania, Latvia, Estonia, Finland, Czech Republic, Serbia, Belarus, Ukraine, Moldova, Georgia, Azerbaijan, Armenia.
Asia: the UAE, Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Uzbekistan, Saudi Arabia, Vietnam, India, Pakistan, Thailand.
North America: Mexico.
South America: Argentina, Brazil.
color,
make,
model,
body style,
vehicle category,
vehicle orientation (front, rear, or side),
special vehicle type (taxi, route transport, carsharing, ambulance, police, rescue service, gas service, military, road service, other special),
vehicle weight and body size.
Note
Special vehicle recognition as well as vehicle weight and body size recognition work for selected countries only and may not work for your country. For more information please contact your manager or our support team (support@ntechlab.com).
Vehicle recognition along with the attributes choice can be configured at the installation level. This section describes how to enable vehicle and vehicle attribute recognition in case this step has been skipped during installation.
To enable recognition of vehicles and their attributes, do the following:
Specify neural network models for vehicle recognition and vehicle attribute recognition in the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file.Important
Be sure to choose the right acceleration type for each model, matching the acceleration type of
findface-extraction-api: CPU or GPU. Be aware thatfindface-extraction-apion CPU can work only with CPU-models, whilefindface-extraction-apion GPU supports both CPU- and GPU-models.Open the
findface-extraction-api.yamlconfiguration file.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
Specify the vehicle detector model in the
detectors -> modelssection by pasting the following code:GPU
detectors: ... models: ... gustav: aliases: - car - efreitor model: detector/car.gustav_accurate.006.gpu.fnk options: min_object_size: 32 resolutions: - 256x256 - 384x384 - 512x512 - 768x768 - 1024x1024 - 1536x1536 - 2048x2048 ...
CPU
detectors: ... models: ... gustav: aliases: - car - efreitor model: detector/car.gustav_accurate.006.cpu.fnk options: min_object_size: 32 resolutions: - 256x256 - 384x384 - 512x512 - 768x768 - 1024x1024 - 1536x1536 - 2048x2048 ...
Make sure that the
objects -> carsection contains thequality_attribute: car_qualityand thebase_normalizer: facenorm/cropbbox.v2.gpu.fnkor thebase_normalizer: facenorm/cropbbox.v2.cpu.fnk, depending on your acceleration type:GPU
objects: ... car: base_normalizer: facenorm/cropbbox.v2.gpu.fnk quality_attribute: car_quality ...
CPU
objects: ... car: base_normalizer: facenorm/cropbbox.v2.cpu.fnk quality_attribute: car_quality ...
Specify the normalizers required for the extractors. If you need license plate recognition, specify the
carlicplatenormalizer. For other extractors, specify thecropbboxnormalizer.Normalizer
Normalizer model
Used for extractors
carlicplate
carnorm/briacon.v0.gpu.fnkcarnorm/briacon.v0.cpu.fnkcar_license_platecropbbox
facenorm/cropbbox.v2.gpu.fnkfacenorm/cropbbox.v2.cpu.fnkcar_license_plate_quality,car_description,car_quality,car_special_types11,car_categories,car_orientation,car_weight_types7GPU
normalizers: ... models: ... carlicplate: model: carnorm/briacon.v0.gpu.fnk ... cropbbox: model: facenorm/cropbbox.v2.gpu.fnk ...
CPU
normalizers: ... models: ... carlicplate: model: carnorm/briacon.v0.cpu.fnk ... cropbbox: model: facenorm/cropbbox.v2.cpu.fnk ...
Specify the extraction models in the
extractors -> modelssection, subject to the extractors you want to enable:GPU
extractors: ... models: car_categories: carattr/carattr.categories.v0.gpu.fnk car_color: '' car_container_number: '' car_description: carattr/description.v0.gpu.fnk car_emben: carrec/bottas.gpu.fnk car_license_plate: carattr/carattr.license_plate.v7.gpu.fnk car_license_plate_quality: carattr/carattr.license_plate_quality.v1.gpu.fnk car_license_plate_visibility: '' car_make: '' car_orientation: carattr/carattr.orientation.v0.gpu.fnk car_quality: carattr/carattr.quality.v0.gpu.fnk car_special_types: '' car_special_types11: carattr/carattr.special_types11.v1.gpu.fnk car_trash: '' car_weight_types7: carattr/carattr.weight_types7.v0.gpu.fnk
CPU
extractors: ... models: car_categories: carattr/carattr.categories.v0.cpu.fnk car_color: '' car_container_number: '' car_description: carattr/description.v0.cpu.fnk car_emben: carrec/bottas.cpu.fnk car_license_plate: carattr/carattr.license_plate.v7.cpu.fnk car_license_plate_quality: carattr/carattr.license_plate_quality.v1.cpu.fnk car_license_plate_visibility: '' car_make: '' car_orientation: carattr/carattr.orientation.v0.cpu.fnk car_quality: carattr/carattr.quality.v0.cpu.fnk car_special_types: '' car_special_types11: carattr/carattr.special_types11.v1.cpu.fnk car_trash: '' car_weight_types7: carattr/carattr.weight_types7.v0.cpu.fnk
The following extractors are available:
Extractor
Configure as follows
vehicle feature vector
car_emben: carrec/bottas.cpu.fnkcar_emben: carrec/bottas.gpu.fnklicense plate number
car_license_plate: carattr/carattr.license_plate.v7.cpu.fnkcar_license_plate_quality: carattr/carattr.license_plate_quality.v1.cpu.fnkcar_license_plate: carattr/carattr.license_plate.v7.gpu.fnkcar_license_plate_quality: carattr/carattr.license_plate_quality.v1.gpu.fnkset of attributes: make / color / model / body style
car_description: carattr/description.v0.cpu.fnkcar_description: carattr/description.v0.gpu.fnkvehicle image quality
car_quality: carattr/carattr.quality.v0.cpu.fnkcar_quality: carattr/carattr.quality.v0.gpu.fnkspecial vehicle
car_special_types: carattr/carattr.special_types11.v1.cpu.fnkcar_special_types: carattr/carattr.special_types11.v1.gpu.fnkvehicle category
car_categories: carattr/carattr.categories.v0.cpu.fnkcar_categories: carattr/carattr.categories.v0.gpu.fnkvehicle weight and body size
car_weight_types7: carattr/carattr.weight_types7.v0.cpu.fnkcar_weight_types7: carattr/carattr.weight_types7.v0.gpu.fnkvehicle orientation
car_orientation: carattr/carattr.orientation.v0.cpu.fnkcar_orientation: carattr/carattr.orientation.v0.gpu.fnkTip
To leave a model disabled, pass the empty value
''to the relevant parameter. Do not remove the parameter itself. Otherwise, the system will be searching for the default model.extractors: ... models: car_categories: '' car_color: '' car_container_number: '' car_description: '' car_emben: '' car_license_plate: '' car_license_plate_quality: '' car_license_plate_visibility: '' car_make: '' car_orientation: '' car_quality: '' car_special_types: '' car_special_types11: '' car_trash: '' car_weight_types7: ''
Restart the
findface-multi-findface-extraction-api-1container.sudo docker container restart findface-multi-findface-extraction-api-1
Modify the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file.In the
modelssection, specify the vehicle detector, normalizer, and extractor models as shown in the example below:GPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: ... detectors: ... car: fnk_path: /usr/share/findface-data/models/detector/car.gustav_normal.004.gpu.fnk min_size: 60 ... normalizers: ... car_norm: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.gpu.fnk car_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.gpu.fnk ... extractors: ... car_quality: fnk_path: /usr/share/findface-data/models/carattr/carattr.quality.v0.gpu.fnk normalizer: car_norm_quality
CPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: ... detectors: ... car: fnk_path: /usr/share/findface-data/models/detector/car.gustav_normal.004.cpu.fnk min_size: 60 ... normalizers: ... car_norm: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk car_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk ... extractors: ... car_quality: fnk_path: /usr/share/findface-data/models/carattr/carattr.quality.v0.cpu.fnk normalizer: car_norm_quality
Make sure that the
objects -> carsection is included:objects: ... car: normalizer: car_norm quality: car_quality track_features: ''
Restart the
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
Open the
/opt/findface-multi/configs/findface-video-manager/findface-video-manager.yamlconfiguration file and make sure it contains thecarsection indetectorsthat looks similar to the example below. Note that thefilter_min_qualityparameter is set to0.65by default. You can increase it to get more accurate results from cameras and other video sources. The best practice is to set this value to0.73.sudo vi /opt/findface-multi/configs/findface-video-manager/findface-video-manager.yaml detectors: ... car: filter_min_quality: 0.65 filter_min_size: 100 filter_max_size: 8192 roi: '' fullframe_crop_rot: false fullframe_use_png: false jpeg_quality: 95 overall_only: true realtime_post_first_immediately: false realtime_post_interval: 1 realtime_post_every_interval: false track_interpolate_bboxes: true track_miss_interval: 1 track_overlap_threshold: 0.25 track_max_duration_frames: 0 track_send_history: false post_best_track_frame: true post_best_track_normalize: true post_first_track_frame: false post_last_track_frame: false tracker_type: simple_iou track_deep_sort_matching_threshold: 0.65 track_deep_sort_filter_unconfirmed_tracks: true track_object_is_principal: false track_history_active_track_miss_interval: 0
Note
Configuration of the
filter_min_qualityparameter affects the results that you get from video sources only and does not affect vehicle counting results and quality of the vehicle images that are used in the records. To get more accurate counting results or vehicle records with better quality images, configure theMINIMUM_CAR_QUALITYparameter individually in thefindface-multi-legacy.pyconfiguration file. This setting will be described below.Enable recognition of vehicles and their attributes in the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. Do the following:In the
FFSECURITYsection, set'ENABLE_CARS': True.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FFSECURITY = { ... # optional objects to detect 'ENABLE_CARS': True, ...
In the same section, you can configure the
MINIMUM_CAR_QUALITYparameter. By default, it’s set to0.65, which is the recommended minimum value for vehicle image quality, sufficient to make a record or use an object in a counter.FFSECURITY = { ... # minimum car quality sufficient to add it to a card and use object in counter # p.s. set this parameter equals to `0.73` for more accurate results 'MINIMUM_CAR_QUALITY': 0.65, # model: [carattr.quality.v0] ...
If you need more accurate counting results or vehicle records with better quality images, you can increase the
MINIMUM_CAR_QUALITYthreshold. The best practice is to set this value to0.73.In the
FFSECURITYsection, specify the attributes you want to display for vehicle recognition events.# available features are: category, description, license_plate, orientation, special_vehicle_type, weight_type 'CAR_EVENTS_FEATURES': ['license_plate', 'category', 'special_vehicle_type', 'description', 'weight_type', 'orientation'],
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
In the web interface, navigate to Video Source. Select a camera in the Cameras tab (or an uploaded file in the Uploads tab, or an external detector in the corresponding tab). Navigate to the General tab. Select Vehicles in the Detectors section.
Interpret Vehicle Recognition Results
Vehicle recognition results can be accessed from the Episodes & Events tab of the web interface.
In certain scenarios vehicle recognition results lack description attributes and their values, or it is displayed that the attribute value is unknown.
The attribute value is unknown. If you see in the recognition results that the attribute value is unknown, it means that the attribute recognition confidence is below the threshold, specified for the attribute in the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. The system will not display the recognized value in this case. Instead, the recognized value will be lost.You can configure the threshold for the vehicle attribute in the
findface-multi-legacy.pyconfiguration file.Important
The above does not apply to the license plate recognition. License plate recognition threshold is set to the minimum value by default. If you see in the recognition results that the license plate number is unknown, it means that it’s hardly or not at all visible on the object. License plate country, region, and color recognition depends on the license plate number recognition results. The values of these attributes will remain unknown if the license plate number is unknown.
Note
License plate region and color are predicted for the United Arab Emirates (UAE) only. The values of these attributes will be marked as unknown for other countries.
Lack of description attributes and their values. If you do not see description attributes (i.e., vehicle make, body, model, color) and their values in the recognition results, it happens for the following reasons:
Recognition of a vehicle category, as well as vehicle weight and body size recognition, is still under development. FindFace Multi accurately recognizes cars of category B and their attributes. When it comes to vehicle weight and body size, recognition of cars and trucks under 3.5 tons and their attributes is supported. However, the system can provide false positive results on other vehicle categories and weight classes and their description attributes. That’s why description attributes and their values are displayed for a car category only, and, when it comes to vehicle weight classes, for cars and trucks under 3.5 tons. For other vehicle categories and weight classes, the description attributes and their recognition values are hidden. The recognition confidence of the description attribute does not matter in this case - it can be below or above the threshold.
The same thing happens if a vehicle category is unknown: description attributes along with recognition values will not be displayed. The recognition confidence of the description attribute is not a determining factor in this case.
Subject to your needs, this behavior can be changed in the
CAR_EVENTS_FEATURES_TO_NULLsection of thefindface-multi-legacy.pyconfiguration file.
Read further to learn how to adjust the threshold for a vehicle attribute or to enable the display of description attributes for some or all vehicle categories and unknown vehicles.
In this section:
Configure the threshold for a vehicle attribute
The default threshold for each vehicle attribute is set to the optimum value. You can change it if necessary.
Open the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file. In the FFSECURITY section, modify threshold values for selected attributes:
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
'CAR_DESCRIPTION_THRESHOLD': {
'make': 0.5,
'model': 0.5,
'body': 0.5,
'color': 0.5,
}, # model: [description.v0]
'SPECIAL_VEHICLE_TYPE_THRESHOLD': 0.5, # model: [special_types11.v1]
'CAR_CATEGORY_THRESHOLD': 0.6, # model: [categories.v0]
'CAR_WEIGHT_TYPE_THRESHOLD': 0.5, # model: [weight-types7.v0]
'CAR_ORIENTATION_THRESHOLD': 0.8, # model [orientation.v0]
...
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker container restart findface-multi-findface-multi-legacy-1
Configure the display of description attributes in the recognition results
Based on your needs, you can enable the display of description attributes for unsupported vehicle categories or unknown vehicles. By default, the display of description attributes is enabled only for cars of the category B and for vehicles of the weight categories B_light (cars under 3.5 tons) and B_heavy (trucks under 3.5 tons). Use information from the tables to match a vehicle and its category or weight class.
Vehicle categories Category
Vehicle
A
motorcycle (including moped), scooter, quad bike
B
car
BE
car with a trailer
C
truck
CE
truck with a trailer
D
bus
DE
articulated bus
other
a vehicle that doesn’t fall within any of the above-listed types
Vehicle weight classes Weight class
Vehicle
B_light
car under 3.5 tons
B_heavy
truck under 3.5 tons
BE
car with a trailer
C_light
truck under 12 tons
C_heavy
truck over 12 tons
D_light
single-deck bus
D_long
articulated bus
other
a vehicle that doesn’t fall within any of the above-listed types
Warning
The default setting is set to the optimum value. Please contact our technical experts (support@ntechlab.com) before making any adjustment.
To enable the display of description attributes for unsupported vehicle categories or unknown vehicles, open the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file. In the FFSECURITY section, find the CAR_EVENTS_FEATURES_TO_NULL setting:
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
'CAR_EVENTS_FEATURES_TO_NULL': {
'category': {
# features, that will be nulled for all categories
'default': ['make', 'body', 'model', 'color'],
'unknown': ['make', 'body', 'model', 'color'],
# categories with other behavior
'B': [],
},
'weight_type': {
# features, that will be nulled for all weight types
'default': ['make', 'body', 'model', 'color'],
# weight types with other behavior
'B_light': [],
'B_heavy': [],
},
},
...
To enable the display of description attributes, remove them from the selected parameter. E.g., if you want to display description attributes for all vehicle categories, except for unknown vehicles, remove them from the default parameter. Note that unknown vehicles are excluded from the default parameter and are configured through the unknown parameter.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
'CAR_EVENTS_FEATURES_TO_NULL': {
'category': {
# features, that will be nulled for all categories
'default': [],
'unknown': ['make', 'body', 'model', 'color'],
# categories with other behavior
'B': [],
},
'weight_type': {
# features, that will be nulled for all weight types
'default': [],
# weight types with other behavior
'B_light': [],
'B_heavy': [],
},
},
...
To enable the display of description attributes for a specific vehicle category or for a vehicle of a certain weight class, add its name to the exception within the categories with other behavior or the weight types with other behavior sections. Use information from the tables Vehicle categories and Vehicle weight classes to match a vehicle and its category or weight class. E.g., if you want to add buses (category D) and single-deck buses (weight class D_light) to the exception, do the following:
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
'CAR_EVENTS_FEATURES_TO_NULL': {
'category': {
# features, that will be nulled for all categories
'default': ['make', 'body', 'model', 'color'],
'unknown': ['make', 'body', 'model', 'color'],
# categories with other behavior
'B': [],
'D': [],
},
'weight_type': {
# features, that will be nulled for all weight types
'default': ['make', 'body', 'model', 'color'],
# weight types with other behavior
'B_light': [],
'B_heavy': [],
'D_light': [],
},
},
...
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker container restart findface-multi-findface-multi-legacy-1
Special Vehicle Recognition
FindFace Multi can recognize a special vehicle and its type. You can set up special vehicle recognition during the installation or enable it later, using the instruction in the Enable Vehicle and Vehicle Attribute Recognition section.
Note
Special vehicle recognition works for selected countries only and may not work for your country. For more information please contact your manager or our support team (support@ntechlab.com).
Recognition of the following vehicle types is supported:
Taxi: taxi, including cargo taxi. In some cases, a cargo taxi can be defined as “other special”.
Public road transport: route transport, i.e. buses, share taxis, trolleybuses.
Carsharing: supported carsharing services are Citydrive, Yandex.Drive, Delimobil, BelkaCar. Carsharing services that look similar to those listed above, may also be recognized.
Ambulance: ambulance, including mobile intensive care units (both white and yellow vehicles).
Police: police, including traffic police cars.
Rescue service: fire service and EMERCOM vehicles.
Gas rescue and emergency services: overall yellow or white cars, including those with contrasting red doors, typical red stripes, and the
04/104sign on board.Military: military vehicles, including the National Guard vehicles.
Road service: road service vehicles, including municipal vehicles, construction equipment, and vehicles of the Center for the Organization of Road Traffic.
Other special: a special vehicle that does not fall within the above-listed types.
Not special: not a special vehicle.
Enable Body and Body Attribute Recognition
FindFace Multi allows you to recognize individual human bodies and body attributes.
The body attributes are as follows:
gender:
male;
female;
age (by group):
0-16 years;
17-35 years;
36-50 years;
50+ years;
clothing type:
generalized category of upper body wear: long sleeves, short sleeves, no sleeve;
specific type of upper body wear: jacket, coat, sleeveless vest, sweatshirt, T-shirt, shirt, dress;
type of lower body wear: pants, skirt, shorts, nondescript;
type of headgear: hat/cap, hood/headscarf, none;
clothing color (top/bottom);
presence of personal protective equipment (PPE):
PPE item: vest, helmet;
PPE color;
PPE recognition score;
whether a person has a bag:
on the back;
in hand(s).
Recognition of human bodies and their attributes can be configured at the installation level. This section describes how to enable body and body attribute recognition in case this step has been skipped during installation.
To enable recognition of human bodies and their attributes, do the following:
Specify neural network models for body object and body attribute recognition in the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file.Important
Be sure to choose the right acceleration type for each model, matching the acceleration type of
findface-extraction-api: CPU or GPU. Be aware thatfindface-extraction-apion CPU can work only with CPU-models, whilefindface-extraction-apion GPU supports both CPU- and GPU-models.Open the
findface-extraction-api.yamlconfiguration file.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
Specify the body detector model in the
detectors -> modelssection by pasting the following code:GPU
detectors: ... models: ... body_gustav: aliases: - body - edie - shiloette - glen model: detector/body.gustav_accurate.019.gpu.fnk options: min_object_size: 32 resolutions: - 256x256 - 384x384 - 512x512 - 768x768 - 1024x1024 - 1536x1536 - 2048x2048 ...
CPU
detectors: ... models: ... body_gustav: aliases: - body - edie - shiloette - glen model: detector/body.gustav_accurate.019.cpu.fnk options: min_object_size: 32 resolutions: - 256x256 - 384x384 - 512x512 - 768x768 - 1024x1024 - 1536x1536 - 2048x2048 ...
Make sure that the
objects -> bodysection contains thequality_attribute: body_qualityand thebase_normalizer: facenorm/cropbbox.v2.gpu.fnkor thebase_normalizer: facenorm/cropbbox.v2.cpu.fnk, depending on your acceleration type:GPU
objects: ... body: base_normalizer: facenorm/cropbbox.v2.gpu.fnk quality_attribute: body_quality ...
CPU
objects: ... body: base_normalizer: facenorm/cropbbox.v2.cpu.fnk quality_attribute: body_quality ...
Make sure that the
normalizerssection contains a model for thecropbboxnormalizer, as shown in the example below. This normalizer is required for the extractors.GPU
normalizers: ... models: ... cropbbox: model: facenorm/cropbbox.v2.gpu.fnk ...
CPU
normalizers: ... models: ... cropbbox: model: facenorm/cropbbox.v2.cpu.fnk ...
Specify the extraction models in the
extractors -> modelssection, subject to the extractors you want to enable:GPU
extractors: ... models: body_action_base6: '' body_action_car: '' body_action_fights: '' body_age_gender: pedattr/pedattr.age_gender.v0.gpu.fnk body_bags: pedattr/pedattr.bags.v0.gpu.fnk body_clothes: pedattr/pedattr.clothes_type.v0.gpu.fnk body_clothes34671: '' body_color: pedattr/pedattr.color.v1.gpu.fnk body_emben: pedrec/pedrec.durga.gpu.fnk body_fall: '' body_handface: '' body_protective_equipment: pedattr/pedattr.protective.v1.gpu.fnk body_quality: pedattr/pedattr.quality.v0.gpu.fnk
CPU
extractors: ... models: body_action_base6: '' body_action_car: '' body_action_fights: '' body_age_gender: pedattr/pedattr.age_gender.v0.cpu.fnk body_bags: pedattr/pedattr.bags.v0.cpu.fnk body_clothes: pedattr/pedattr.clothes_type.v0.cpu.fnk body_clothes34671: '' body_color: pedattr/pedattr.color.v1.cpu.fnk body_emben: pedrec/pedrec.durga.cpu.fnk body_fall: '' body_handface: '' body_protective_equipment: pedattr/pedattr.protective.v1.cpu.fnk body_quality: pedattr/pedattr.quality.v0.cpu.fnk
The following extractors are available:
Extractor
Configure as follows
age and gender
body_age_gender: pedattr/pedattr.age_gender.v0.gpu.fnkbody_age_gender: pedattr/pedattr.age_gender.v0.cpu.fnkpresence of bag
body_bags: pedattr/pedattr.bags.v0.gpu.fnkbody_bags: pedattr/pedattr.bags.v0.cpu.fnkclothing type
body_clothes: pedattr/pedattr.clothes_type.v0.gpu.fnkbody_clothes: pedattr/pedattr.clothes_type.v0.cpu.fnkclothing color
body_color: pedattr/pedattr.color.v1.gpu.fnkbody_color: pedattr/pedattr.color.v1.cpu.fnkindividual body feature vector
body_emben: pedrec/pedrec.durga.gpu.fnkbody_emben: pedrec/pedrec.durga.cpu.fnkpresence of protective equipment
body_protective_equipment: pedattr/pedattr.protective.v1.gpu.fnkbody_protective_equipment: pedattr/pedattr.protective.v1.cpu.fnkbody quality
body_quality: pedattr/pedattr.quality.v0.gpu.fnkbody_quality: pedattr/pedattr.quality.v0.cpu.fnkTip
To leave a model disabled, pass the empty value
''to the relevant parameter. Do not remove the parameter itself. Otherwise, the system will be searching for the default model.Important
For body recognition to work properly, the
body_embenand thebody_qualityextractors must be enabled.GPU
extractors: ... models: body_action_base6: '' body_action_car: '' body_action_fights: '' body_age_gender: '' body_bags: '' body_clothes: '' body_clothes34671: '' body_color: '' body_emben: pedrec/pedrec.durga.gpu.fnk body_fall: '' body_handface: '' body_protective_equipment: '' body_quality: pedattr/pedattr.quality.v0.gpu.fnk
CPU
extractors: ... models: body_action_base6: '' body_action_car: '' body_action_fights: '' body_age_gender: '' body_bags: '' body_clothes: '' body_clothes34671: '' body_color: '' body_emben: pedrec/pedrec.durga.cpu.fnk body_fall: '' body_handface: '' body_protective_equipment: '' body_quality: pedattr/pedattr.quality.v0.cpu.fnk
Restart the
findface-multi-findface-extraction-api-1container.sudo docker container restart findface-multi-findface-extraction-api-1
Modify the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file.In the
modelssection, specify the body neural network models by analogy with the example below:GPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: ... detectors: ... body: fnk_path: /usr/share/findface-data/models/detector/body.jasmine_fast.018.gpu.fnk min_size: 60 ... normalizers: ... body_norm: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.gpu.fnk body_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.gpu.fnk ... extractors: ... body_quality: fnk_path: /usr/share/findface-data/models/pedattr/pedattr.quality.v0.gpu.fnk normalizer: body_norm_quality
CPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml models: ... detectors: ... body: fnk_path: /usr/share/findface-data/models/detector/body.jasmine_fast.018.cpu.fnk min_size: 60 ... normalizers: ... body_norm: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk body_norm_quality: fnk_path: /usr/share/findface-data/models/facenorm/cropbbox.v2.cpu.fnk ... extractors: ... body_quality: fnk_path: /usr/share/findface-data/models/pedattr/pedattr.quality.v0.cpu.fnk normalizer: body_norm_quality
Make sure that the
objects -> bodysection is included:objects: ... body: normalizer: body_norm quality: body_quality track_features: ''
Restart the
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
Open the
/opt/findface-multi/configs/findface-video-manager/findface-video-manager.yamlconfiguration file and make sure it contains thebodysection indetectorsthat looks similar to the example below.sudo vi /opt/findface-multi/configs/findface-video-manager/findface-video-manager.yaml detectors: ... body: filter_min_quality: 0.6 filter_min_size: 70 filter_max_size: 8192 roi: '' fullframe_crop_rot: false fullframe_use_png: false jpeg_quality: 95 overall_only: true realtime_post_first_immediately: false realtime_post_interval: 1 realtime_post_every_interval: false track_interpolate_bboxes: true track_miss_interval: 1 track_overlap_threshold: 0.25 track_max_duration_frames: 0 track_send_history: false post_best_track_frame: true post_best_track_normalize: true post_first_track_frame: false post_last_track_frame: false tracker_type: simple_iou track_deep_sort_matching_threshold: 0.65 track_deep_sort_filter_unconfirmed_tracks: true track_object_is_principal: false track_history_active_track_miss_interval: 0
Enable recognition of bodies and body attributes in the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. Do the following:In the
FFSECURITYsection, set'ENABLE_BODIES': True.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FFSECURITY = { ... # optional objects to detect 'ENABLE_BODIES': True, ...
To improve quality of body recognition, we recommend that you enable additional attribute analysis. In this case, the system compares not only the feature vectors of two bodies but also their attributes. A conclusion about the bodies’ match is only made if both the feature vectors and attributes of the bodies coincide.
You can use the following attributes for additional analysis:
bottom_color: color of lower body wear;top_color: color of upper body wear;headwear: type and absence/presence of headgear;detailed_upper_clothes: specific type of upper body wear, e.g., jacket;upper_clothes: generalized category of upper body wear: long sleeves, short sleeves, no sleeve;lower_clothes: type of lower body wear, e.g., pants;helmet_type: helmet type by color, visibility, absence/presence;vest_type: vest type by color, visibility, absence/presence;age_group: belonging to any of four age groups: 0-16, 17-35, 36-50, 50+ years;gender: male or female.
To enable additional attribute analysis, set
Truein theFFSECURITY->EXTRA_BODY_MATCHINGsection for the attributes that you want to compare. Setmin_confidencevalue between 0 and 1.FFSECURITY = { # use additional features for extra confidence when matching body by emben 'EXTRA_BODY_MATCHING': { 'bottom_color': {'enabled': False, 'min_confidence': 0}, 'top_color': {'enabled': False, 'min_confidence': 0}, 'headwear': {'enabled': False, 'min_confidence': 0}, 'detailed_upper_clothes': {'enabled': False, 'min_confidence': 0}, 'upper_clothes': {'enabled': False, 'min_confidence': 0}, 'lower_clothes': {'enabled': False, 'min_confidence': 0}, 'helmet_type': {'enabled': False, 'min_confidence': 0}, 'vest_type': {'enabled': False, 'min_confidence': 0}, 'age_group': {'enabled': False, 'min_confidence': 0}, 'gender': {'enabled': False, 'min_confidence': 0}, },
Note
Contact our technical experts (support@ntechlab.com) for advice on
min_confidenceoptimum value.If you decide that you do not need additional attribute analysis, then skip this configuration and proceed to the next step.
In the
FFSECURITYsection, specify the body attributes that you want to display for the body recognition events.# available features: age_gender, bags, clothes, color, protective_equipment 'BODY_EVENTS_FEATURES': ['protective_equipment', 'age_gender', 'bags', 'color', 'clothes'],
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
In the web interface, navigate to Video Source. Select a camera in the Cameras tab (or an uploaded file in the Uploads tab, or an external detector in the corresponding tab). Navigate to the General tab. Select Bodies in the Detectors section.
Enable Face Liveness Detection
The FindFace Multi face liveness detector tells apart live faces from face representations, such as face images, videos, or masks. The liveness detector estimates face liveness with a certain level of confidence and returns the confidence score along with a binary result real/fake, depending on the pre-defined liveness threshold.
The face liveness detector can be automatically enabled and configured during the installation. If you skipped this step, you can manually enable it later, following the instructions below.
Note
The face liveness detector functions on both GPU- and CPU-acceleration. However, it is much slower on CPU.
In this section:
Enable Face Liveness Detector
To enable the face liveness detector, do the following:
Open the
/opt/findface-multi/configs/findface-video-worker/findface-video-worker.yamlconfiguration file. In thelivenesssection, specify the neural network models as shown in the example:GPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml liveness: fnk: /usr/share/findface-data/models/faceattr/liveness.pacs.v2.gpu.fnk norm: /usr/share/findface-data/models/facenorm/facenorm.multicrop_full_crop2x_size400.gpu.fnk ...
CPU
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml liveness: fnk: /usr/share/findface-data/models/faceattr/liveness.pacs.v2.cpu.fnk norm: /usr/share/findface-data/models/facenorm/facenorm.multicrop_full_crop2x_size400.cpu.fnk ...
Restart the
findface-multi-findface-video-worker-1container.sudo docker container restart findface-multi-findface-video-worker-1
Configure Liveness Threshold
If necessary, you can adjust the liveness threshold in the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file. The liveness detector will estimate face liveness with a certain level of confidence. Depending on the threshold value, it will return a binary result real or fake.
Note
The default value is optimal. Before changing the threshold, we recommend that you seek advice from our experts at support@ntechlab.com.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
# feature specific confidence thresholds
'LIVENESS_THRESHOLD': 0.674, # model: [liveness.pacs.v2]
...
Restart the findface-multi-findface-multi-legacy-1 container if you have configured the liveness threshold.
sudo docker container restart findface-multi-findface-multi-legacy-1
Migrate Feature Vectors to a Different Neural Network Model
Tip
Do not hesitate to contact our experts on migration by support@ntechlab.com.
Important
In case, you are doing migration as part of the system update to a newer version, complete the update first. Only after that can you proceed to the migration.
This section is about how to migrate object feature vectors to another neural network model.
Do the following:
Create a backup of the Tarantool-based feature vector database in any directory of your choice, for example,
/etc/ffmulti_dump.sudo docker exec -it findface-multi-findface-sf-api-1 bash -c "mkdir ffmulti_dump; cd ffmulti_dump && /storage-api-dump -config /etc/findface-sf-api.ini" sudo docker cp findface-multi-findface-sf-api-1:/ffmulti_dump /etc
Create new shards that will host regenerated feature vectors.
Navigate to the
/opt/findface-multi/data/findface-tarantool-serverdirectory and find out the number of shards by counting the number of directories.Note
There are eight shards in the example below.
cd /opt/findface-multi/data/findface-tarantool-server ls -l shard-001 shard-002 shard-003 shard-004 shard-005 shard-006 shard-007 shard-008
Create directories that will host files of the new shards.
sudo mkdir -p shard-01{1..8}/{index,snapshots,xlogs}
Open the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file and replace the extraction models with the new ones in thebody_emben,car_emben, andface_embenparameters, depending on the object types you want to migrate.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml extractors: ... models: ... body_emben: pedrec/<new_model_body>.cpu<gpu>.fnk ... car_emben: carrec/<new_model_car>.cpu<gpu>.fnk ... face_emben: face/<new_model_face>.cpu<gpu>.fnk ...
Restart the
findface-multi-findface-extraction-api-1container.cd /opt/findface-multi/ sudo docker-compose restart findface-extraction-api
In the
docker-compose.yamlfile, create a new service for each new shard. To do that, copy the existing service and replace the name of the shard, ports inCFG_LISTEN_PORT,TT_LISTENand the path to the shard involumessection.sudo vi docker-compose.yaml services: ... findface-tarantool-server-shard-**011**: depends_on: [findface-ntls] environment: {CFG_EXTRA_LUA: loadfile("/tnt_schema.lua")(), CFG_LISTEN_HOST: 127.0.0.1, CFG_LISTEN_PORT: '8111', CFG_NTLS: '127.0.0.1:3133', TT_CHECKPOINT_COUNT: 3, TT_CHECKPOINT_INTERVAL: '14400', TT_FORCE_RECOVERY: 'true', TT_LISTEN: '127.0.0.1:32011', TT_MEMTX_DIR: snapshots, TT_MEMTX_MEMORY: '2147483648', TT_WAL_DIR: xlogs, TT_WORK_DIR: /var/lib/tarantool/FindFace} image: docker.int.ntl/ntech/universe/tntapi:ffserver-9.230407.1 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./data/findface-tarantool-server/shard-**011**:/var/lib/tarantool/FindFace', './configs/findface-tarantool-server/tnt-schema.lua:/tnt_schema.lua:ro'] ...
Start the new shards by upping the containers.
sudo docker-compose up -d
Create a configuration file with migration settings
migration.yamlbased on the example below.sudo vi migration.yaml extraction-api: timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 0s extraction-api: http://127.0.0.1:18666 storage-api-from: # current location of the gallery timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 10s max-idle-conns-per-host: 20 shards: - master: http://127.0.0.1:8101/v2/ slave: '' - master: http://127.0.0.1:8102/v2/ slave: '' - master: http://127.0.0.1:8103/v2/ slave: '' - master: http://127.0.0.1:8104/v2/ slave: '' - master: http://127.0.0.1:8105/v2/ slave: '' - master: http://127.0.0.1:8106/v2/ slave: '' - master: http://127.0.0.1:8107/v2/ slave: '' - master: http://127.0.0.1:8108/v2/ slave: '' storage-api-to: timeouts: connect: 5s response_header: 30s overall: 35s idle_connection: 10s max-idle-conns-per-host: 20 shards: - master: http://127.0.0.1:8111/v2/ slave: '' - master: http://127.0.0.1:8112/v2/ slave: '' - master: http://127.0.0.1:8113/v2/ slave: '' - master: http://127.0.0.1:8114/v2/ slave: '' - master: http://127.0.0.1:8115/v2/ slave: '' - master: http://127.0.0.1:8116/v2/ slave: '' - master: http://127.0.0.1:8117/v2/ slave: '' - master: http://127.0.0.1:8118/v2/ slave: '' workers_num: 3 objects_limit: 1000 extraction_batch_size: 8 normalized_storage: type: webdav enabled: True webdav: upload-url: http://127.0.0.1:3333/uploads/ s3: endpoint: '' bucket-name: '' access-key: '' secret-access-key: '' secure: False region: '' public-url: '' operation-timeout: 30
In the
storage-api-fromsection, specify the old shards to migrate the data from.storage-api-from: # current location of the gallery ... shards: - master: http://127.0.0.1:8101/v2/ slave: '' - master: http://127.0.0.1:8102/v2/ slave: '' - master: http://127.0.0.1:8103/v2/ slave: '' - master: http://127.0.0.1:8104/v2/ slave: '' - master: http://127.0.0.1:8105/v2/ slave: '' - master: http://127.0.0.1:8106/v2/ slave: '' - master: http://127.0.0.1:8107/v2/ slave: '' - master: http://127.0.0.1:8108/v2/ slave: '' ...
In the
storage-api-tosection, specify the new shards that will host migrated data.storage-api-to: ... shards: - master: http://127.0.0.1:8111/v2/ slave: '' - master: http://127.0.0.1:8112/v2/ slave: '' - master: http://127.0.0.1:8113/v2/ slave: '' - master: http://127.0.0.1:8114/v2/ slave: '' - master: http://127.0.0.1:8115/v2/ slave: '' - master: http://127.0.0.1:8116/v2/ slave: '' - master: http://127.0.0.1:8117/v2/ slave: '' - master: http://127.0.0.1:8118/v2/ slave: '' ...
Copy the
migration.yamlfile into thefindface-multi-findface-sf-api-1container. Launch thesf-api-migrateutility with the-configoption and provide themigration.yamlconfiguration file.sudo docker cp migration.yaml findface-multi-findface-sf-api-1:/ sudo docker exec findface-multi-findface-sf-api-1 ./sf-api-migrate -config migration.yaml
Note
The migration process can take up a significant amount of time if there are many events and records in the system.
After the migration is complete, remove services for the old shards from the
docker-compose.yamlfile and stop their containers.sudo docker-compose up -d --remove-orphans
Open the
/opt/findface-multi/configs/findface-sf-api/findface-sf-api.yamlconfiguration file and adjust the shards ports, subject to the new shards settings. Restart thefindface-multi-findface-sf-api-1container.sudo vi /opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml storage-api: shards: - master: http://127.0.0.1:8111/v2/ slave: '' - master: http://127.0.0.1:8112/v2/ slave: '' - master: http://127.0.0.1:8113/v2/ slave: '' - master: http://127.0.0.1:8114/v2/ slave: '' - master: http://127.0.0.1:8115/v2/ slave: '' - master: http://127.0.0.1:8116/v2/ slave: '' - master: http://127.0.0.1:8117/v2/ slave: '' - master: http://127.0.0.1:8118/v2/ slave: '' sudo docker-compose restart findface-sf-api
Migrate clusters as well if this functionality is enabled in your system. To do so, execute the following command:
Note
List the object types to migrate as the command options:
--face,--body,--car.sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py migrate_clusters --face --body --car --use-best-event --use-thumbnail --force-clustering
As a result, the system will regenerate feature vectors for the existing cluster events and automatically launch the scheduled clustering to rebuild clusters.
Deactivate findface-liveness-api installed with FindFace Multi
If you don’t utilize the findface-liveness-api service installed with FindFace Multi and the face-based authentication does not apply to your system, we recommend deactivating findface-liveness-api.
Do the following:
Stop the
findface-multi-findface-liveness-api-1container and disable its autostart by executing:sudo docker container stop findface-multi-findface-liveness-api-1 sudo docker update --restart=no findface-multi-findface-liveness-api-1
Open the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlconfiguration file.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
Disable the neural network models used by the
findface-liveness-apiservice. To do so, pass the empty value''to themulticrop_full_crop2xandface_livenessparameters.Note
Do not remove the parameter itself. Otherwise, the system will be searching for the default model.
normalizers: models: ... multicrop_full_crop2x: '' ... extractors: models: ... face_liveness: '' ...
Restart the
findface-multi-findface-extraction-api-1container.sudo docker container restart findface-multi-findface-extraction-api-1
Maintenance and Troubleshooting
Update to FindFace Multi 2.1
Tip
If you use our product FindFace Security deployed on Ubuntu 18.04, upgrade it to FindFace Multi 1.2 and then update it to FindFace Multi 2.1.
If you use FindFace Multi 2.0, update it to FindFace Multi 2.1 by following the instruction.
The Incidents and VNS plugins, operable in FindFace CIBR, are not supported in FindFace Multi 2.1. That’s why there’s no need to enable ffsecurity_incidents, ffsecurity_vns, and ffsecurity_puppeteer plugins in the findface-multi-legacy.py configuration file after product update.
Integration with Axxon Next is included into external VMS integration in FindFace Multi 2.1 and configured through VMS integration plugin.
If you created person and car cards with custom fields in FindFace Multi 1.2 and want them to be displayed in the FindFace Multi 2.1 UI, then it is necessary to copy the CUSTOM_FIELDS section from the old configuration file in step #2 and paste it to the new configuration file in step #8.
In this section:
Update FindFace Multi 1.2 to FindFace Multi 2.1
Important
Before you start the update, disable all cameras.
To update FindFace Multi 1.2 to FindFace Multi 2.1, do the following:
Create a backup copy of the old schema of the Tarantool-based feature vector database:
sudo cp /etc/findface-security/tnt_schema.lua /etc/findface-security/old_tnt_schema.lua
Starting from version 2.0, one of the most significant differences between FindFace Multi and earlier versions of the product is the structure of the Tarantool biometric database (so called meta-schema). The new structure is created as a set of spaces, while in previous versions of the product there was only one space by default in the structure of the Tarantool-based database.
Open the
/etc/findface-security/config.pyconfiguration file. Save the values of the following parameters for later use:EXTERNAL_ADDRESS,SECRET_KEY,VIDEO_DETECTOR_TOKEN,ROUTER_URL,CUSTOM_FIELDS.sudo vi /etc/findface-security/config.py EXTERNAL_ADDRESS = "http://172.20.77.58" ... # use pwgen -sncy 50 1|tr "'" "." to generate your own unique key SECRET_KEY = 'c8b533847bbf7142102de1349d33a1f6' FFSECURITY = { 'VIDEO_DETECTOR_TOKEN': '381b0f4a20495227d04185ab02f5085f', ... 'ROUTER_URL': 'http://172.20.77.58', ... # -- Custom model fields -- # Edit CUSTOM_FIELDS -> `human_card` section to customize human card fields. # Edit CUSTOM_FIELDS -> `face_object` section to customize face object fields. # Below is an example with every field type possible. # 'CUSTOM_FIELDS': { # 'human_card': { # 'items': [ # { # 'name': 'personid', # 'default': '', # 'label': 'PersonID', # 'display': ['list', 'form'], # 'description': 'Sigur person ID', # 'editable': False # }, # { # 'name': 'firstname', # 'default': '', # 'label': 'First Name', # 'display': ['list', 'form'], # 'description': 'Sigur first name', # 'editable': False # }, # { # 'name': 'lastname', # 'default': '', # 'label': 'Last Name', # 'display': ['list', 'form'], # 'description': 'Sigur last name', # 'editable': False # }, # { # 'name': 'version', # 'default': '', # 'label': 'Version', # 'display': ['list', 'form'], # 'description': 'Sigur photo version', # 'editable': False # } # ], # 'filters': [ # { # 'name': 'personid', # 'label': 'Sigur person ID filter', # 'field': 'personid' # } # ] # }, # 'face_object': { # 'items': [ # { # "field_name": "tag_name_1", # "type": "string", # "default": "change_me" # }, # { # "field_name": "tag_name_2", # "type": "uint", # "default": 123 # }, # { # "field_name": "tag_name_3", # "type": "bool", # "default": True # }, # ] # } # }, }
Stop the
findface-securityservice.sudo systemctl stop findface-security.service
Create a backup of the Tarantool-based feature vector database in any directory of your choice, for example,
/etc/findface_dump.sudo mkdir -p /etc/findface_dump cd /etc/findface_dump sudo findface-storage-api-dump -config /etc/findface-sf-api.ini
To avoid port conflicts, stop and disable all services before installing a new version.
Note
There are eight shards in the example below. If it differs with the number of shards in your system, adjust the below command accordingly. E.g., for the system with sixteen shards, replace
tarantool@shard-00{1..8}.servicewithtarantool@shard-0{01..16}.service. Get the list of active shards with thels /etc/tarantool/instances.enabled/command.You may stop and disable the services one by one:
sudo systemctl stop postgresql.service sudo systemctl stop postgresql@10-main sudo systemctl stop findface-*.service sudo systemctl stop pgbouncer.service sudo systemctl stop tarantool@shard-00{1..8}.service sudo systemctl stop nats-server.service sudo systemctl stop etcd.service sudo systemctl stop mongod.service sudo systemctl stop mongodb.service sudo systemctl stop memcached.service sudo systemctl stop nginx.service sudo systemctl disable postgresql.service sudo systemctl disable postgresql@10-main sudo systemctl disable pgbouncer.service sudo systemctl disable findface-extraction-api.service sudo systemctl disable findface-security.service sudo systemctl disable findface-security-onvif.service sudo systemctl disable findface-sf-api.service sudo systemctl disable findface-ntls.service sudo systemctl disable findface-video-manager.service sudo systemctl disable findface-video-worker-cpu.service sudo systemctl disable findface-video-worker-gpu.service sudo systemctl disable findface-counter.service sudo systemctl disable findface-liveness-api.service sudo systemctl disable findface-video-streamer-cpu.service sudo systemctl disable findface-video-streamer-gpu.service sudo systemctl disable findface-video-storage.service sudo systemctl disable tarantool@shard-00{1..8}.service sudo systemctl disable nats-server.service sudo systemctl disable etcd.service sudo systemctl disable mongod.service sudo systemctl disable mongodb.service sudo systemctl disable memcached.service sudo systemctl disable nginx.service
Or you may use the following compact commands instead:
sudo systemctl stop postgresql.service postgresql@10-main findface-*.service pgbouncer.service tarantool@shard-00{1..8}.service nats-server.service etcd.service mongod.service mongodb.service memcached.service nginx.service sudo systemctl disable postgresql.service postgresql@10-main pgbouncer.service findface-extraction-api.service findface-security.service findface-security-onvif.service findface-sf-api.service findface-ntls.service findface-video-manager.service findface-video-worker-cpu.service findface-video-worker-gpu.service findface-counter.service findface-liveness-api.service findface-video-streamer-cpu.service findface-video-streamer-gpu.service findface-video-storage.service tarantool@shard-00{1..8}.service nats-server.service etcd.service mongod.service mongodb.service memcached.service nginx.service
Install the FindFace Multi 2.1 instance. Don’t forget to prepare a server first:
See:
After FindFace Multi installation, stop all the containers from the
/opt/findface-multi/directory.cd /opt/findface-multi/ sudo docker-compose stop
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file and paste saved on step #2 values for the parametersEXTERNAL_ADDRESS,SECRET_KEY,VIDEO_DETECTOR_TOKEN,ROUTER_URL, andCUSTOM_FIELDSinto it.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py ... # Use pwgen -sncy 50 1|tr "'" "." to generate your own unique key SECRET_KEY = '002231ccb690586f4d33e98322c591bb' ... SERVICE_EXTERNAL_ADDRESS = 'http://172.20.77.58' # EXTERNAL_ADDRESS is used to access objects created inside FFSecurity via external links. EXTERNAL_ADDRESS = 'http://172.20.77.58' ... # findface-video-worker authorization token 'VIDEO_DETECTOR_TOKEN': '8977e1b0067d43f6c908d0bf60363255', ... # findface-video-worker face posting address, # it must be set to either FFSecurity EXTERNAL_ADDRESS (by default) # or findface-facerouter url (in some specific cases) 'ROUTER_URL': 'http://127.0.0.1:80',
Open the old version of the
findface-ntlsconfiguration file available at/etc/findface-ntls.cfgand check it against the new version/opt/findface-multi/configs/findface-ntls/findface-ntls.yaml. Move all the custom parameters from the old version to the new one. Do the same for other components, e.g., forfindface-extract-api, check/etc/findface-extract-api.iniagainst/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml, forfindface-sf-api, check/etc/findface-sf-api.iniagainst/opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml, etc.sudo vi /etc/findface-ntls.cfg sudo vi /opt/findface-multi/configs/findface-ntls/findface-ntls.yaml sudo vi /etc/findface-extraction-api.ini sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml sudo vi /etc/findface-sf-api.ini sudo vi /opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml
Important
Follow these rules to transfer parameters from the old configuration file to the new one:
If there is a new neural network model in the new configuration file, replace it with the previous one (considering that the previous model is still included in FindFace Multi 2.1), and if the previous model is missing in FindFace Multi 2.1, then do not change anything. In this case, you will have to do migration to a different neural network model.
If a parameter had an empty value in the old configuration file, but has a certain value in the new configuration file, delete its value in the new configuration file.
Keep as is those parameters that were not included in the old configuration file, but are present in the new configuration file.
Modify the Tarantool database structure by applying the
tnt_schema.luaschema from FindFace Multi 2.1.sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 make-tnt-schema | sudo tee /etc/findface-security/tnt_schema.lua
Purge data from all the directories relevant to active shards.
sudo rm /opt/ntech/var/lib/tarantool/shard-*/{index,snapshots,xlogs}/*
Copy the meta-schema of the
defaultspace from theold_tnt_schema.luaconfiguration file to the newtnt_schema.luaconfiguration file, so that the old meta-schema is still available. An easy way to do it is to follow these steps:12.1. In the
/etc/findface-security/old_tnt_schema.luafile, rename the following fields:meta_scheme --> meta_scheme_default meta_indexes --> meta_indexes_default
12.2. In the new configuration file
/etc/findface-security/tnt_schema.lua, replace the following lines at the beginning of the file:cfg_spaces = { default = { meta_scheme = { -- internal.normalized_id: { default = '', field_type = 'string', id = 1, name = 'normalized_id', }, -- internal.feat: { default = '', field_type = 'string', id = 2, name = 'feat', }, }, meta_indexes = {} },
with these ones:
dofile("/etc/findface-security/old_tnt_schema.lua") spaces = { default = { meta_scheme=meta_scheme_default, meta_indexes=meta_indexes_default },
Navigate to the directory with Tarantool configuration file(s)
/etc/tarantool/instances.available/. Import new meta-schematnt_schema.luainto each configuration fileshard-00*.lua, as in the example below.sudo vi /etc/tarantool/instances.available/shard-00*.lua dofile("/etc/findface-security/tnt_schema.lua") FindFace = require("FindFace") FindFace.start("127.0.0.1", 8104, { license_ntls_server="127.0.0.1:3133", replication = replication_master, spaces = spaces })
Remove the default configuration file
FindFace.lua, generated by thefindface-tarantool-serverpackage, as it will block the restart, required on the next step.sudo rm -rf /etc/tarantool/instances*/FindFace.lua
Restart the
findface-tarantool-servershards.TNT=$(ls /etc/tarantool/instances.enabled/ | cut -c 7,8,9) for i in $TNT; do sudo systemctl restart tarantool@shard-$i.service ; done
Upon completion of the above steps, the shards will still keep the old galleries created within the default space, but new spaces (e.g., ffsec_body_objects_space, ffsec_face_clusters_space, and so on) will also become available.
Restore old data from the backup. The data will be restored as it existed previously: all galleries will stay within the
defaultspace.sudo systemctl start findface-ntls.service cd /etc/findface_dump for x in *.json; do sudo findface-storage-api-restore -config /etc/findface-sf-api.ini < "$x"; done
Migrate galleries from the
defaultspace to new spaces:sudo systemctl start findface-sf-api.service sudo systemctl start nginx.service sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 /opt/findface-security/bin/python3 /tigre_prototype/manage.py migrate_tnt_space
Perform PostgreSQL database migrations for FindFace Multi 2.1 compatibility. Do the following:
18.1. Navigate to the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. In theDATABASES -> defaultsection, temporarily replacePASSWORDwith the old one, used in the/etc/findface-security/config.pyconfiguration file.Important
Make sure to write down the password from the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. You will need it later.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py DATABASES = { 'default': { 'ENGINE': 'django.db.backends.postgresql', 'DISABLE_SERVER_SIDE_CURSORS': True, 'NAME': 'ffsecurity', 'HOST': '127.0.0.1', 'PORT': 5439, 'USER': 'ntech', 'PASSWORD': 'XXXXXXXXXXXXXXXX' } }
18.2. In the
/etc/pgbouncer/pgbouncer.inifile, add the following line to thedatabasessection:ffsecurity_session = dbname=ffsecurity host=localhost port=5432 user=ntech pool_mode=session pool_size=10
18.3. On the host system, perform the database migration:
sudo systemctl start postgresql.service sudo systemctl start pgbouncer.service sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 /opt/findface-security/bin/python3 /tigre_prototype/manage.py migrate
18.4. Back up an existing database with PostgreSQL, installed on the host system.
cd /opt/findface-multi/ sudo -u postgres pg_dump --verbose --disable-triggers ffsecurity | sudo tee dump_ffsecurity.sql sudo -u postgres pg_dump -t auth_group -t ffsecurity_adgroupguid -t ffsecurity_deviceblacklistrecord -t ffsecurity_ffsecauthsession -t ffsecurity_grouppermission -t ffsecurity_runtimesetting -t ffsecurity_user -t ffsecurity_user_groups -t ffsecurity_user_user_permissions -t ffsecurity_userkeyvalue -t knox_authtoken -t ffsecurity_watchlistpermission -t ffsecurity_cameragrouppermission --data-only --verbose --no-acl --no-owner --disable-triggers ffsecurity | sudo tee dump_identity_provider.sql
18.5. Back up role permissions.
sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 /opt/findface-security/bin/python3 /tigre_prototype/manage.py dump_permissions | sudo tee permissions.csv
18.6. Change back the password, replaced in step #18.1
18.7. Stop all the services.
sudo systemctl stop findface-sf-api.service nginx.service tarantool@shard-00{1..8}.service postgresql.service pgbouncer.service
18.8. Open the
/opt/findface-multi/docker-compose.yamlfile and copyPOSTGRES_PASSWORDvalue to use it in further commands.18.9. Recreate the
ffsecuritydatabase to clean it up from the default data. Paste{POSTGRES_PASSWORD}value that you previously copied in step #18.8 into the command below:sudo docker-compose up -d postgresql sudo docker exec -it -u postgres findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} dropdb ffsecurity" sudo docker exec -it -u postgres findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} createdb -O ntech --encoding=UTF-8 --lc-collate=C.UTF-8 --lc-ctype=C.UTF-8 --template=template0 ffsecurity"
18.10. Restore data into the recreated
ffsecuritydatabase. Paste{POSTGRES_PASSWORD}value that you previously copied in step #18.8 into the command below:sudo docker exec -i findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} psql --username postgres ffsecurity" < dump_ffsecurity.sql
18.11. Recreate the
ffsecurity_identity_providerdatabase to clean it up from the default data. Paste{POSTGRES_PASSWORD}value that you previously copied in step #18.8 into the command below:sudo docker exec -it -u postgres findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} dropdb ffsecurity_identity_provider" sudo docker exec -it -u postgres findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} createdb -O ntech --encoding=UTF-8 --lc-collate=C.UTF-8 --lc-ctype=C.UTF-8 --template=template0 ffsecurity_identity_provider"
18.12. Run migration.
sudo docker-compose up -d pgbouncer sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/identity-provider:ffmulti-2.1.0 /opt/findface-security/bin/python3 /tigre_prototype/manage.py migrate
18.13. Restore data into the recreated
ffsecurity_identity_providerdatabase. Paste{POSTGRES_PASSWORD}value that you previously copied in step #18.8 into the command below:sudo docker exec -i findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} psql --username postgres ffsecurity_identity_provider" < dump_identity_provider.sql
18.14. Start all the services.
sudo docker-compose up -d
18.15. Run the command to create records in the outbox table for watch lists and camera groups. Paste
{POSTGRES_PASSWORD}value that you previously copied in step #18.8 into the command below:sudo docker exec -i findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} pg_dump --username postgres -f cg_wl_permissions.sql -t ffsecurity_cameragrouppermission -t ffsecurity_watchlistpermission --data-only ffsecurity_identity_provider" sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 /opt/findface-security/bin/python3 /tigre_prototype/manage.py create_outbox sudo docker exec -i findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} psql --username postgres -c 'TRUNCATE ffsecurity_cameragrouppermission, ffsecurity_watchlistpermission RESTART IDENTITY;' ffsecurity_identity_provider" sudo docker exec -i findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} psql --username postgres ffsecurity_identity_provider < cg_wl_permissions.sql"
18.16. Before you restore role permissions into the
identity_providerservice, examine the/opt/findface-multi/permissions.csvfile. Make sure to replace*_ffsecauthtokenwith*_authtokenif any. This is mostly applicable to those cases when FindFace Multi 1.2 installation was an upgrade from earlier versions of the product.After that, restore role permissions into the
identity_providerservice.sudo docker run --rm --network host -v /opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.py:/etc/findface-security/config.py:ro -v $(pwd)/permissions.csv:/var/permissions.csv:ro docker.int.ntl/ntech/multi/multi/identity-provider:ffmulti-2.1.0 /opt/findface-security/bin/python3 /tigre_prototype/manage.py load_permissions /var/permissions.csv
18.17. Copy full frame photos, normalized images, and the license file. Copy files from the
/opt/ntech/license/folder into the/opt/findface-multi/data/findface-ntls/folder, from the/var/lib/findface-security/uploads/folder into the/opt/findface-multi/data/findface-multi-legacy/uploads/folder, from the/var/lib/ffupload/uploads/folder into the/opt/findface-multi/data/findface-upload/uploads/folder.sudo cp -r /opt/ntech/license/* /opt/findface-multi/data/findface-ntls/ sudo cp -r /var/lib/findface-security/uploads/* /opt/findface-multi/data/findface-multi-legacy/uploads/ sudo cp -r /var/lib/ffupload/uploads/* /opt/findface-multi/data/findface-upload/uploads/ sudo chmod 777 -R /opt/findface-multi/data/findface-upload/uploads/ sudo chown www-data:www-data -R /opt/findface-multi/data/findface-upload/uploads/*
Alternatively, the above folders can be directly mounted into the relevant docker containers via the
docker-compose.yamlfile, like in the example below:sudo vi /opt/findface-multi/docker-compose.yaml findface-upload: image: docker.int.ntl/ntech/universe/upload:ffserver-9.230407.1 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-upload/40-ffupload.sh:/docker-entrypoint.d/40-ffupload.sh:ro', '/var/lib/ffupload:/var/lib/ffupload'] findface-multi-identity-provider: depends_on: [pgbouncer, nats, findface-sf-api, findface-liveness-api, etcd] environment: {ADMIN_PASSWORD: <ADMIN_PASSWORD>} image: docker.int.ntl/ntech/multi/multi/identity-provider:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-identity-provider/findface-multi-identity-provider.py:/etc/findface-security/config.py:ro', '/var/lib/findface-security/uploads:/var/lib/findface-security/uploads'] findface-multi-legacy: depends_on: [pgbouncer, nats, findface-sf-api, findface-counter, findface-liveness-api, etcd] environment: {ADMIN_PASSWORD: <ADMIN_PASSWORD>} image: docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro', '/var/lib/findface-security/uploads:/var/lib/findface-security/uploads'] findface-multi-ui: depends_on: [findface-multi-legacy] image: docker.int.ntl/ntech/multi/multi/ui:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-ui/nginx-site.conf:/etc/nginx/conf.d/default.conf:ro', '/var/lib/findface-security/uploads:/var/lib/findface-security/uploads'] cleaner: command: [run-cleaner-service] depends_on: [pgbouncer, findface-sf-api, findface-multi-legacy] image: docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro', '/var/lib/findface-security/uploads:/var/lib/findface-security/uploads'] findface-multi-file-mover: command: [python3, /app/service.py, --config=/etc/config.yml] depends_on: [findface-multi-legacy, nats-jetstream] image: docker.int.ntl/ntech/multi/multi/file-mover:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-file-mover/findface-multi-file-mover.yaml:/etc/config.yml:ro', '/var/lib/findface-security/uploads:/mnt/slow_storage/uploads']
18.18. To move Tarantool data, do the following:
Stop all FindFace Multi containers:
sudo docker-compose down
Start the old shards and the
findface-sf-apiservice again:sudo systemctl start tarantool@shard-00{1..8}.service findface-sf-api.service
Create a new backup of the feature vector database:
sudo mkdir -p /etc/findface_dump_final sudo findface-storage-api-dump -output-dir=/etc/findface_dump_final -config /etc/findface-sf-api.ini
Stop the rest of the services, clear the
instances.enableddirectory, start the containers again, and perform the storage-api-restore operation:sudo systemctl stop tarantool@shard-00{1..8}.service findface-sf-api.service findface-ntls.service sudo rm /etc/tarantool/instances.enabled/* sudo docker-compose up -d sudo findface-storage-api-restore -config /opt/findface-multi/configs/findface-sf-api/findface-sf-api.yaml /etc/findface_dump_final/*.json
The update has been completed, but the new version includes new neural network models, so it is also necessary to migrate feature vectors to a different neural network model, or you can use old neural network models by moving them from the /usr/share/findface-data/models/ directory to the /opt/findface-multi/models/ directory, specifying them in the /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml file.
Note that object detection models in FindFace Multi 1.2 are stored in separate directories, i.e., face detection models are stored in the /usr/share/findface-data/models/facedet/ directory, car and body detection models are stored in the /usr/share/findface-data/models/cadet/ and /usr/share/findface-data/models/pedet/ directories respectively. In FindFace Multi 2.1 all object detection models can be found in the /opt/findface-multi/models/detector/ directory. When moving old neural network models from the /usr/share/findface-data/models/ directory to the /opt/findface-multi/models/ directory, make sure to place all object detection models (facedet, cadet, pedet) to the /opt/findface-multi/models/detector/ directory.
See for reference:
$ ls -lash /usr/share/findface-data/models
total 52K
4.0K drwxr-xr-x 13 root root 4.0K Jul 15 14:48 .
4.0K drwxr-xr-x 3 root root 4.0K Jul 15 14:48 ..
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 cadet
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 carattr
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 carnorm
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 carrec
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 face
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 faceattr
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 facedet
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 facenorm
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 pedattr
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 pedet
4.0K drwxr-xr-x 2 root root 4.0K Jul 15 14:48 pedrec
$ ls -lash /opt/findface-multi/models/
total 44K
4.0K drwxr-xr-x 11 root root 4.0K Jul 17 13:37 .
4.0K drwxr-xr-x 6 root root 4.0K Jul 19 15:31 ..
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 13:37 carattr
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 13:37 carnorm
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 13:36 carrec
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 16:20 detector
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 16:24 face
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 16:24 faceattr
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 13:37 facenorm
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 13:37 pedattr
4.0K drwxr-xr-x 2 root root 4.0K Jul 17 13:37 pedrec
Important
We highly recommend disabling the Ubuntu automatic update to preserve the FindFace Multi compatibility with the installation environment. In this case, you will be able to update your OS manually, fully controlling which packages to update.
To disable the Ubuntu automatic update, execute the following commands:
sudo apt-get remove unattended-upgrades
sudo systemctl stop apt-daily.timer
sudo systemctl disable apt-daily.timer
sudo systemctl disable apt-daily.service
sudo systemctl daemon-reload
Update FindFace Multi 2.0 / 2.0.1 to FindFace Multi 2.1
No matter whether you are updating from FindFace Multi 2.0 or 2.0.1 to FindFace Multi 2.1, the update process will be the same. Do the following:
In a user home directory, create a folder for FindFace Multi 2.0.1 backup data.
mkdir ~/201/
Back up FindFace Multi configuration files and the
docker-compose.yamlfile and switch to the/opt/findface-multi/directory.sudo tar -cvzf ~/201/configs.tar.gz -C /opt/findface-multi/ configs sudo cp /opt/findface-multi/docker-compose.yaml ~/201/ cd /opt/findface-multi/
Back up PostgreSQL data.
sudo docker-compose cp postgresql:/bitnami/postgresql ~/201/pg_bkp
Back up Tarantool data.
sudo docker exec -it findface-multi-findface-sf-api-1 bash -c "mkdir ffmulti_dump; cd ffmulti_dump && /storage-api-dump -config /etc/findface-sf-api.ini" sudo docker cp findface-multi-findface-sf-api-1:/ffmulti_dump ~/201/
List content of the
201/folder, located in a user home directory. It will look like this.ls ~/201/ configs.tar.gz docker-compose.yaml ffmulti_dump pg_bkp
Stop all FindFace Multi containers.
sudo docker-compose down
Remove the
/opt/findface-multi/data/findface-tarantool-server/and/opt/findface-multi/data/postgresql/folders.sudo rm -r /opt/findface-multi/data/findface-tarantool-server/ sudo rm -r /opt/findface-multi/data/postgresql/
Install the FindFace Multi 2.1 instance into the
/opt/findface-multidirectory.Restore Tarantool data from the backup.
Copy the
~/201/ffmulti_dump/folder to thefindface-multi-findface-sf-api-1container.sudo docker cp ~/201/ffmulti_dump/ findface-multi-findface-sf-api-1:/
Restore Tarantool data.
sudo docker exec -it findface-multi-findface-sf-api-1 bash -c 'cd ffmulti_dump && for x in *.json; do /storage-api-restore -config /etc/findface-sf-api.ini < "$x"; done;'
In the new configuration files, replace the
NTECH USERpassword with the one, used in FindFace Multi 2.0.1. The old password can be obtained from theconfigs/postgresql/40-init.sqlfile in the backup archiveconfigs.tar.gz. TheNTECH USERpassword should be replaced in the following configuration files:findface-multi-legacy.pyfindface-multi-line-crossing-analytics.yamlfindface-multi-identity-provider.pyfindface-multi-audit.pypgbouncer -> userlist.txtpostgresql -> 40-init.sqlfindface-counter.yaml
In the new configuration file
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py, replace theSECRET_KEYvalue in theGENERAL SETTINGSsection and'VIDEO_DETECTOR_TOKEN'value in theFINDFACE SECURITY SETTINGSsection with the ones, used in FindFace Multi 2.0.1. The old parameter values can be obtained from theconfigs/findface-multi-legacy/findface-multi-legacy.pyfile in the backup archiveconfigs.tar.gz.Compare previous configuration files against the new ones. Move all custom parameters from the previous version to the new one. Compare configuration files for the services
findface-ntls,findface-extract-api,findface-sf-api, etc. You may need to apply changes to thedocker-compose.yamlfile if you previously configured it to bring or exclude services or made any other changes.Important
Follow these rules to transfer parameters from the old configuration file to the new one:
If there is a new neural network model in the new configuration file, replace it with the previous one (considering that the previous model is still included in FindFace Multi 2.1), and if the previous model is missing in FindFace Multi 2.1, then do not change anything. In this case, you will have to do migration to a different neural network model.
If a parameter had an empty value in the old configuration file, but has a certain value in the new configuration file, delete its value in the new configuration file.
Keep as is those parameters that were not included in the old configuration file, but are present in the new configuration file.
There are fresh neural network models for body recognition and vehicle recognition in FindFace Multi 2.1.
In the
/opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yamlfile, replace thepedrec.cliomodel with thepedrec.durgamodel for the extractorbody_emben.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
CPU
extractors: ... models: ... body_emben: pedrec/pedrec.durga.cpu.fnk ...
GPU
extractors: ... models: ... body_emben: pedrec/pedrec.durga.gpu.fnk ...
Also, in the same configuration file, replace the
alonsomodel with thebottasmodel for the extractorcar_emben.sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
CPU
extractors: ... models: ... car_emben: carrec/bottas.cpu.fnk ...
GPU
extractors: ... models: ... car_emben: carrec/bottas.gpu.fnk ...
Stop all FindFace Multi containers from the
/opt/findface-multidirectory.cd /opt/findface-multi sudo docker-compose down
Copy data from the
~/201/pg_bkp/data/directory to the/opt/findface-multi/data/postgresqldirectory.sudo cp -r ~/201/pg_bkp/data/ /opt/findface-multi/data/postgresql
Start all FindFace Multi containers.
sudo docker-compose up -d
Note
You can additionally deploy Video Recorder. See Deploy Video Recorder Step-by-Step.
Back Up and Recover FindFace Multi and Its Data
You can back up FindFace Multi before uninstalling it to recover the product and its data later on.
In this section:
Back up FindFace Multi
To back up your FindFace Multi instance and its data, run the following commands:
sudo tar -cvzf ~/configs.tar.gz -C /opt/findface-multi/ configs
sudo tar -cvzf ~/data.tar.gz -C /opt/findface-multi/ data
sudo cp /opt/findface-multi/docker-compose.yaml ~/
If you use tiered data storage and have enabled saving to hot storage, then as a result of the backup, thumbnails and full frame images that were not transferred to the cold storage will not be saved. If you need to back up all images, do the following:
Install the
rsynkutility.sudo apt install rsync
Force the transfer of images by copying them from the hot to the cold storage.
Note
The
ssd-pathis a path (directory) to the SSD that is being used for the hot storage. Here and on replace thessd-pathwith an existing SSD directory.rsync -aP /ssd-path/event/ /opt/findface-multi/data/findface-multi-legacy/uploads/event/
Back up your FindFace Multi instance and its data by running commands above.
Recover FindFace Multi and Its Data
To restore FindFace Multi and its data from the backup, do the following:
Download the installer file
findface-*.run.Put the
.runfile into some directory on the designated host (for example,/home/username).From this directory, make the
.runfile executable.Note
Be sure to specify the actual file name instead of
findface-*.chmod +x findface-*.run
Execute the
.runfile.sudo ./findface-*.run
Go through the installation process as described here.
After you have finished the installation, to restore FindFace Multi, its data and configuration files, stop all FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose stop
Remove new configuration files and data generated and created by the installer and restore them from the backup.
Important
Compare your old
~/docker-compose.yamlfile that you have copied during the backup process against the new one/opt/findface-multi/docker-compose.yaml. You may need to apply changes to the old~/docker-compose.yamlfile if you previously configured it to bring or exclude services or made any other changes. You must be fully aware of what you are doing by replacing the newdocker-compose.yamlfile with the old one.sudo rm -r /opt/findface-multi/configs/* sudo tar -xvf ~/configs.tar.gz -C /opt/findface-multi/ sudo rm -r /opt/findface-multi/data/* sudo tar -xvf ~/data.tar.gz -C /opt/findface-multi/ sudo cp ~/docker-compose.yaml /opt/findface-multi/
Restart FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose up -d
Modify Feature Vector Database Structure
Sometimes it may be necessary to apply a new structural schema to your Tarantool-based feature vector database, for example, when updating to the latest version of the product, or when you want to enhance the default database structure with additional parameters, advanced face metadata, and so on.
In this section:
About Database Structure
In FindFace Multi, the database structure is set via the /opt/findface-multi/configs/findface-tarantool-server/tnt-schema.lua file.
The structure is created as a set of spaces and fields. Each field is described with the following parameters:
id: field id;name: field name, must be the same as the name of a relevant object parameter;field_type: data type (unsigned|string|set[string]|set[unsigned]);default: field default value. If a default value exceeds1e14 – 1, use a string data type to specify it, for example,"123123.."instead of123123...
You can find the default tnt-schema.lua file here.
Structure Modification
To modify the database structure, do the following:
Create a backup of the Tarantool-based feature vector database in any directory of your choice, for example,
/etc/ffmulti_dump.sudo docker exec -it findface-multi-findface-sf-api-1 bash -c "mkdir ffmulti_dump; cd ffmulti_dump && /storage-api-dump -config /etc/findface-sf-api.ini" sudo docker cp findface-multi-findface-sf-api-1:/ffmulti_dump /etc
Modify the database structure by applying the new schema to the
tnt-schema.luafile.sudo docker run --rm --network host --volume '/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro' docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 make-tnt-schema | sudo tee /opt/findface-multi/configs/findface-tarantool-server/tnt-schema.lua
Open the
docker-compose.yamlfile. Make sure each Tarantool shard service contains defined variableCFG_EXTRA_LUA: loadfile("/tnt_schema.lua")in theenvironmentsection.sudo vi /opt/findface-multi/docker-compose.yaml ... findface-tarantool-server-shard-001: depends_on: [findface-ntls] environment: {CFG_EXTRA_LUA: loadfile("/tnt_schema.lua")()... ...
Purge data from all the directories relevant to active shards.
sudo rm /opt/findface-multi/data/findface-tarantool-server/shard-*/{index,snapshots,xlogs}/*
Restart the
findface-tarantool-servershards.docker restart findface-multi-findface-tarantool-server-shard-001-1 findface-multi-findface-tarantool-server-shard-002-1 findface-multi-findface-tarantool-server-shard-003-1 findface-multi-findface-tarantool-server-shard-004-1 findface-multi-findface-tarantool-server-shard-005-1 findface-multi-findface-tarantool-server-shard-006-1 findface-multi-findface-tarantool-server-shard-007-1 findface-multi-findface-tarantool-server-shard-008-1
Restore the Tarantool database from the backup.
sudo docker exec -it findface-multi-findface-sf-api-1 bash -c 'cd ffmulti_dump && for x in *.json; do /storage-api-restore -config /etc/findface-sf-api.ini < "$x"; done;'
Important
If some fields were removed from the new database structure, you have to first manually delete the corresponding data from the backup copy.
See also
Check Component Status
Check the status of containers once you have encountered a system problem, using the following commands:
docker ps
sudo docker container inspect <container_id>/<container_name>
sudo docker container stats <container_id>/<container_name>
Logging
Consulting logs is one of the first things you should do to identify a cause of a problem. By default, the FindFace Multi processes are logged to Docker container logs, which can be accessed via the docker logs and docker service logs commands. In addition, Docker uses the json-file logging driver, which caches container logs in JSON. You can configure Docker to use another logging driver, choosing from the multiple logging mechanisms available.
This section describes how to set up Docker to use the journald logging driver, which sends container logs to the systemd journal. In this case, log entries are retrieved using the journalctl command, through the journal API, or the docker logs command. You can configure the systemd journal as well by using the instructions below.
In this section:
Configure Journald
To configure the systemd-journal service, do the following:
Check whether the
/var/log/journaldirectory already exists. If not, create it by executing the following command:sudo mkdir /var/log/journal sudo chmod 2755 /var/log/journal
Open the
/etc/systemd/journald.confconfiguration file. Enable savingjournaldlogs to your hard drive by uncommenting theStorageparameter and changing its value topersistent. Disable filtering insystemd-journalas well:sudo vi /etc/systemd/journald.conf [Journal] ... Storage=persistent ... RateLimitInterval=0 RateLimitBurst=0 ...
If necessary, uncomment and edit the
SystemMaxUseparameter. This parameter determines the maximum volume of log files on your hard drive. Specify its value in bytes or use K, M, G, T, P, E as units for the specified size (equal to \(1024\), \(1024^2\), … bytes).... SystemMaxUse=3G
Restart the
journaldservice.sudo systemctl restart systemd-journald.service
Enabling Journald Logging Driver
To enable Docker to use the journald logging driver, do the following:
Add the following line to the
/etc/docker/daemon.jsonconfiguration file.Tip
This file may not be present on your system. Check whether it exists and if it does not, create the
/etc/dockerdirectory first and then the file.sudo mkdir /etc/docker sudo touch /etc/docker/daemon.json
sudo vi /etc/docker/daemon.json { "log-driver": "journald" }
Stop all Docker containers.
sudo docker stop $(sudo docker ps -a -q)
Restart the Docker service.
sudo systemctl restart docker
Start all Docker containers.
sudo docker start $(sudo docker ps -a -q)
Consult Logs
Use any convenient method to work with the journald logs. The following commands are a good place to start:
Display all logs:
journalctl -fa
Display logs by container ID:
journalctl CONTAINER_ID=<container_id> -f
Display logs by container name:
journalctl CONTAINER_NAME=<container-name> -f
See also
Troubleshoot Licensing and findface-ntls
When troubleshooting licensing and findface-ntls (see Licensing), the first step is to retrieve the licensing information and findface-ntls status. You can do so by sending an API request to findface-ntls. Necessary actions are then to be undertaken, subject to the response content.
Tip
Please do not hesitate to contact our experts on troubleshooting by support@ntechlab.com.
Note
The online licensing is done via the NtechLab Global License Manager license.ntechlab.com. Check its availability. A stable internet connection and DNS are required.
To retrieve the FindFace licensing information and findface-ntls status, execute on the findface-ntls host console:
curl http://localhost:3185/license.json -s | jq
The response will be given in JSON. One of the most significant parameters is last_updated. It indicates in seconds how long ago the local license has been checked for the last time.
Interpret the last_updated value as follows:
[0, 5] — everything is alright.
(5, 30] — there may be some problems with connection, or with the local drive where the license file is stored.
(30; 120] — almost certainly something bad happened.
(120; ∞) — the licensing source response has been timed out. Take action.
"valid" -> "value": false: connection with the licensing source was never established.
curl http://localhost:3185/license.json -s | jq
{
"name": "NTLS",
"time": 1565186356,
"type": "online",
"license_id": "61063ce4b86945e1b70c3bdbedea453b",
"generated": 1514467939,
"last_updated": 5,
"valid": {
"value": true,
"description": ""
},
"source": "/opt/ntech/license/import_b68d7b7ec9a7310d18832035318cff0c9ddf11e3a9ab0ae962fbe48645e196d1.lic",
"limits": [
{
"type": "time",
"name": "end",
"value": 1609161621
},
{
"type": "number",
"name": "faces",
"value": 9007199254740991,
"current": 0
},
{
"type": "number",
"name": "cameras",
"value": 4294967295,
"current": 0
},
{
"type": "number",
"name": "extraction_api",
"value": 256,
"current": 0
},
{
"type": "boolean",
"name": "gender",
"value": true
},
{
"type": "boolean",
"name": "age",
"value": true
},
{
"type": "boolean",
"name": "emotions",
"value": true
},
{
"type": "boolean",
"name": "fast-index",
"value": true
},
{
"type": "boolean",
"name": "sec-genetec",
"value": false
},
{
"type": "boolean",
"name": "beard",
"value": false
},
{
"type": "boolean",
"name": "glasses",
"value": false
},
{
"type": "boolean",
"name": "liveness",
"value": false
}
],
"services": [
{
"name": "video-worker",
"ip": "127.0.0.1:53276"
},
{
"name": "FindFace-tarantool",
"ip": "127.0.0.1:53284"
},
{
"name": "FindFace-tarantool",
"ip": "127.0.0.1:53288"
}
]
}
Manually Purge Old Data from Database
Tip
To schedule automatic database cleanup, see General Settings.
To manually remove old data from the FindFace Multi database, use the cleanup utility. You can separately remove the following data:
matched events (faces, bodies, vehicles) and related episodes,
unmatched events (faces, bodies, vehicles) and related episodes,
full frames of matched events (faces, bodies, vehicles),
full frames of unmatched events (faces, bodies, vehicles),
counter screenshots,
cluster events (faces, bodies, vehicles),
audit-logs,
external VMS events.
The cleanup utility runs in the findface-multi-findface-multi-legacy-1 container. To invoke the cleanup help message, execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --help
usage: manage.py cleanup [-h] [--as-configured]
[--face-events-max-fullframe-unmatched-age FACE_EVENTS_MAX_FULLFRAME_UNMATCHED_AGE]
[--face-events-max-fullframe-matched-age FACE_EVENTS_MAX_FULLFRAME_MATCHED_AGE]
[--face-events-max-unmatched-age FACE_EVENTS_MAX_UNMATCHED_AGE]
[--face-events-max-matched-age FACE_EVENTS_MAX_MATCHED_AGE]
[--body-events-max-fullframe-unmatched-age BODY_EVENTS_MAX_FULLFRAME_UNMATCHED_AGE]
[--body-events-max-fullframe-matched-age BODY_EVENTS_MAX_FULLFRAME_MATCHED_AGE]
[--body-events-max-unmatched-age BODY_EVENTS_MAX_UNMATCHED_AGE]
[--body-events-max-matched-age BODY_EVENTS_MAX_MATCHED_AGE]
[--car-events-max-fullframe-unmatched-age CAR_EVENTS_MAX_FULLFRAME_UNMATCHED_AGE]
[--car-events-max-fullframe-matched-age CAR_EVENTS_MAX_FULLFRAME_MATCHED_AGE]
[--car-events-max-unmatched-age CAR_EVENTS_MAX_UNMATCHED_AGE]
[--car-events-max-matched-age CAR_EVENTS_MAX_MATCHED_AGE]
[--car-cluster-events-max-age CAR_CLUSTER_EVENTS_MAX_AGE]
[--body-cluster-events-max-age BODY_CLUSTER_EVENTS_MAX_AGE]
[--face-cluster-events-max-age FACE_CLUSTER_EVENTS_MAX_AGE]
[--car-cluster-events-keep-best-max-age CAR_CLUSTER_EVENTS_KEEP_BEST_MAX_AGE]
[--body-cluster-events-keep-best-max-age BODY_CLUSTER_EVENTS_KEEP_BEST_MAX_AGE]
[--face-cluster-events-keep-best-max-age FACE_CLUSTER_EVENTS_KEEP_BEST_MAX_AGE]
[--area-activations-max-age AREA_ACTIVATIONS_MAX_AGE]
[--audit-logs-max-age AUDIT_LOGS_MAX_AGE]
[--counter-records-max-age COUNTER_RECORDS_MAX_AGE]
[--external-vms-events-max-age EXTERNAL_VMS_EVENTS_MAX_AGE]
[--external-vms-send-events-status-max-age EXTERNAL_VMS_SEND_EVENTS_STATUS_MAX_AGE]
[--remote-monitoring-events-max-age REMOTE_MONITORING_EVENTS_MAX_AGE]
[--configuration CONFIGURATION] [--version]
[-v {0,1,2,3}] [--settings SETTINGS]
[--pythonpath PYTHONPATH] [--traceback] [--no-color]
[--force-color] [--skip-checks]
Delete FFSecurity entities
optional arguments:
-h, --help show this help message and exit
--as-configured Apply config age options for events, counter records
and clusters. Can't be used with other arguments.
--face-events-max-fullframe-unmatched-age FACE_EVENTS_MAX_FULLFRAME_UNMATCHED_AGE
face events max fullframe unmatched age to clean up
(in seconds)
--face-events-max-fullframe-matched-age FACE_EVENTS_MAX_FULLFRAME_MATCHED_AGE
face events max fullframe matched age to clean up (in
seconds)
--face-events-max-unmatched-age FACE_EVENTS_MAX_UNMATCHED_AGE
face events max unmatched age to clean up (in seconds)
--face-events-max-matched-age FACE_EVENTS_MAX_MATCHED_AGE
face events max matched age to clean up (in seconds)
--body-events-max-fullframe-unmatched-age BODY_EVENTS_MAX_FULLFRAME_UNMATCHED_AGE
body events max fullframe unmatched age to clean up
(in seconds)
--body-events-max-fullframe-matched-age BODY_EVENTS_MAX_FULLFRAME_MATCHED_AGE
body events max fullframe matched age to clean up (in
seconds)
--body-events-max-unmatched-age BODY_EVENTS_MAX_UNMATCHED_AGE
body events max unmatched age to clean up (in seconds)
--body-events-max-matched-age BODY_EVENTS_MAX_MATCHED_AGE
body events max matched age to clean up (in seconds)
--car-events-max-fullframe-unmatched-age CAR_EVENTS_MAX_FULLFRAME_UNMATCHED_AGE
car events max fullframe unmatched age to clean up (in
seconds)
--car-events-max-fullframe-matched-age CAR_EVENTS_MAX_FULLFRAME_MATCHED_AGE
car events max fullframe matched age to clean up (in
seconds)
--car-events-max-unmatched-age CAR_EVENTS_MAX_UNMATCHED_AGE
car events max unmatched age to clean up (in seconds)
--car-events-max-matched-age CAR_EVENTS_MAX_MATCHED_AGE
car events max matched age to clean up (in seconds)
--car-cluster-events-max-age CAR_CLUSTER_EVENTS_MAX_AGE
car cluster events max age to clean up (in seconds)
--body-cluster-events-max-age BODY_CLUSTER_EVENTS_MAX_AGE
body cluster events max age to clean up (in seconds)
--face-cluster-events-max-age FACE_CLUSTER_EVENTS_MAX_AGE
face cluster events max age to clean up (in seconds)
--car-cluster-events-keep-best-max-age CAR_CLUSTER_EVENTS_KEEP_BEST_MAX_AGE
car cluster events keep best max age to clean up (in
seconds)
--body-cluster-events-keep-best-max-age BODY_CLUSTER_EVENTS_KEEP_BEST_MAX_AGE
body cluster events keep best max age to clean up (in
seconds)
--face-cluster-events-keep-best-max-age FACE_CLUSTER_EVENTS_KEEP_BEST_MAX_AGE
face cluster events keep best max age to clean up (in
seconds)
--area-activations-max-age AREA_ACTIVATIONS_MAX_AGE
area activations max age to clean up (in seconds)
--audit-logs-max-age AUDIT_LOGS_MAX_AGE
audit logs max age to clean up (in seconds)
--counter-records-max-age COUNTER_RECORDS_MAX_AGE
counter records max age to clean up (in seconds)
--external-vms-events-max-age EXTERNAL_VMS_EVENTS_MAX_AGE
external vms events max age to clean up (in seconds)
--external-vms-send-events-status-max-age EXTERNAL_VMS_SEND_EVENTS_STATUS_MAX_AGE
external vms send events status max age to clean up
(in seconds)
--remote-monitoring-events-max-age REMOTE_MONITORING_EVENTS_MAX_AGE
remote monitoring events max age to clean up (in
seconds)
--configuration CONFIGURATION
The name of the configuration class to load, e.g.
"Development". If this isn't provided, the
DJANGO_CONFIGURATION environment variable will be
used.
--version show program's version number and exit
-v {0,1,2,3}, --verbosity {0,1,2,3}
Verbosity level; 0=minimal output, 1=normal output,
2=verbose output, 3=very verbose output
--settings SETTINGS The Python path to a settings module, e.g.
"myproject.settings.main". If this isn't provided, the
DJANGO_SETTINGS_MODULE environment variable will be
used.
--pythonpath PYTHONPATH
A directory to add to the Python path, e.g.
"/home/djangoprojects/myproject".
--traceback Raise on CommandError exceptions
--no-color Don't colorize the command output.
--force-color Force colorization of the command output.
--skip-checks Skip system checks.
To entirely remove events and episodes older than a given number of seconds, use the --*-events-max-matched-age/--*-events-max-unmatched-age options, subject to the object type. For example, to remove unmatched car events older than 5 days (432000 seconds), execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --car-events-max-unmatched-age 432000
To remove only matched car events older than 5 days (432000 seconds), execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --car-events-max-matched-age 432000
The following commands remove only full frames of matched/unmatched body events:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --body-events-max-fullframe-matched-age 432000
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --body-events-max-fullframe-unmatched-age 432000
To remove only counter screenshots, execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --counter-records-max-age 432000
To remove only cluster events with faces, execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --face-cluster-events-max-age 432000
To remove only audit logs, execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py cleanup --audit-logs-max-age 432000
Important
You must provide at least one of the mentioned arguments.
Reset Password
To reset a user password to the FindFace Multi web interface, execute the following command:
sudo docker exec -it findface-multi-findface-multi-identity-provider-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py changepassword <username>
S3-Compatible Storage Configuration
S3 (Simple Storage Service) provides reliable and long-term storage of an unlimited number of files and data. It gives a way to avoid the limitations of the file system when it comes to large amounts of data.
Configure S3-Compatible Storage
In this section, we will describe how to configure S3-compatible MinIO storage for operation with FindFace Multi. If you prefer another S3-compatible storage, you are yourself responsible for its configuration.
-
Note
MinIO operation has been tested on Ubuntu only. Refer to the official MinIO documentation to find out whether other OS are supported.
To start a container with MinIO, execute the following command, specify the
MINIO_ROOT_USERandMINIO_ROOT_PASSWORDlogin credentials:docker run -d \ -p 9003:9000 \ -p 9004:9001 \ -e "MINIO_ROOT_USER=admin" \ -e "MINIO_ROOT_PASSWORD=admin_admin" \ -v /path/to/storage_directory:/data \ quay.io/minio/minio server /data --console-address ":9001"
where
/path/to/storage_directoryis a directory, allocated for the S3 storage.MinIO web interface address will depend on the interface that you have selected during FindFace Multi installation in step #4.5. The port in use is
9004, e.g.,192.168.112.254:9004. Log in to the MinIO web interface withMINIO_ROOT_USER,MINIO_ROOT_PASSWORDlogin credentials that you have specified in the previous step.In the MinIO UI, create a bucket (Buckets -> Create Bucket). Assign a default name
ffsec.storageto the bucket.Note
We advise keeping the bucket name
ffsec.storageas suggested by default, because it corresponds to the'STORAGE_BUCKET_NAME'parameter in the/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyand/opt/findface-multi/configs/findface-multi-legacy/findface-multi-identity-provider.pyconfiguration files. If you use another name for a bucket, make sure to adjust it in the above-mentioned configuration files accordingly.
Click on the bucket to open its settings.
Change access policy for the bucket to
publicfor the correct processing of video archives.
In the User -> Access Keys section, create an access key. Write down the
Access keyandSecret keyvalues as you will need them later for FindFace Multi configuration.
Note
MinIO UI may differ. In the earlier versions, set up an access key in the Identity -> Service Accounts -> Create service account section.
Configure FindFace Multi to Employ S3-Compatible Storage
In the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file, locate theEXTERNAL_STORAGE_CONFIGsection.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py EXTERNAL_STORAGE_CONFIG = { 'STORAGE_TYPE': None, 'STORAGE_BUCKET_NAME': 'ffsec.storage', 'ACCESS_KEY_ID': 'access-key', 'SECRET__KEY': 'secret-key', 'REGION_NAME': 'eu-west-3', 'LOCAL_S3_SETTINGS': { 'CONNECTION_ADDRESS': 'localhost:9003', }, }
Configure the
EXTERNAL_STORAGE_CONFIGparameters:The
'STORAGE_TYPE'parameter may take values'LOCAL','AWS',None:None– local file system (default)'LOCAL'– any S3-compatible storage (e.g., MinIO)'AWS'– Amazon S3 storage.
Change the
'STORAGE_TYPE'parameter to'LOCAL'to enable S3-compatible storage.The values of the
'STORAGE_BUCKET_NAME','ACCESS_KEY_ID', and'SECRET__KEY'parameters should correspond to those that were specified during MinIO initialization.'REGION_NAME'– the default value iseu-west-3. It can be changed in MinIO settings.In the
'LOCAL_S3_SETTINGS'->'CONNECTION_ADDRESS', replace thelocalhostwith MinIO IP host address, e.g.,192.168.112.254.
Configure the
/opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.pyfile in the same way. Locate theEXTERNAL_STORAGE_CONFIGsection and apply changes, analogous to those made in the/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyfile in the previous step.sudo vi /opt/findface-multi/configs/findface-multi-identity-provider/findface-multi-identity-provider.py EXTERNAL_STORAGE_CONFIG = { 'STORAGE_TYPE': None, 'STORAGE_BUCKET_NAME': 'ffsec.storage', 'ACCESS_KEY_ID': 'access-key', 'SECRET__KEY': 'secret-key', 'REGION_NAME': 'eu-west-3', 'LOCAL_S3_SETTINGS': { 'CONNECTION_ADDRESS': 'localhost:9003', }, }
In order for the UI to load content (thumbnails, full frame images, and so on), configure the
/opt/findface-multi/configs/findface-multi-ui/nginx-site.conffile. In theserversection, locate theadd_header Content-Security-Policyand add{ip_host_s3}:9003/address, e.g.:sudo vi /opt/findface-multi/configs/findface-multi-ui/nginx-site.conf ... server { listen 80 default_server; listen [::]:80 default_server; ... location / { ... } add_header Content-Security-Policy "default-src 'self' 192.168.112.254:9003/ ; img-src 'self' 192.168.112.254:9003/ blob: data:; media-src 'self' blob:; style-src 'self' 'unsafe-inline';"; ... }
Rebuild all FindFace Multi containers from the
/opt/findface-multidirectory .cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
To make sure that the
FindFace Local S3 Storageservice is running, explore thefindface-multi-findface-multi-legacy-1logs:sudo docker logs findface-multi-findface-multi-legacy-1
Data Transfer Command
To transfer data to either remote or local storage, use the following command:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py transfer_data --to_local / --to_remote
The argument
--to_remotetransfers data to remote storage.The argument
--to_localtransfers data to the local file system.
Before executing the above command, make sure that you have configured FindFace Multi to use S3-compatible storage.
Tiered Data Storage
At peak loads, a HDD does not cope with timely saving of event files. As a result, new events are queued and, at a certain point, the saving process terminates.
This problem can be solved with the help of tiered data storage. It works as follows: first, FindFace Multi saves event images to the hot storage (a size-limited SSD), and then the file moving service transfers them to the cold storage (ordinary HDD). Only thumbnails and full frame images are saved to the hot storage. The remaining data, including videos, record images, reports, and so on, is saved immediately to the cold storage and does not need to be moved. To start using the file moving service, enable the hot storage.
Enable Hot Storage
By default, saving to the hot storage is disabled. The file moving service findface-multi-file-mover starts along with other services but stays non-functional.
To enable the hot storage service, do the following:
Allocate the SSD for the hot storage. Consider the SSD volume: it should be large enough to store the images at peak loads before they are moved to the cold storage.
In the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyfile, configure theFAST_MEDIA_ROOTparameter. By default, it’s set toNone.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FAST_MEDIA_ROOT = '/var/lib/fast-storage/uploads'
Open the
/opt/findface-multi/docker-compose.yamlconfiguration file.In the
/opt/findface-multi/docker-compose.yamlfile, locate thefindface-multi-file-moverservice.findface-multi-file-mover: command: [python3, /app/service.py, --config=/etc/config.yml] depends_on: [findface-multi-legacy, nats-jetstream] image: docker.int.ntl/ntech/multi/multi/file-mover:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-file-mover/findface-multi-file-mover.yaml:/etc/config.yml:ro', './data/findface-multi-legacy/uploads:/mnt/slow_storage/uploads']
Configure the
findface-multi-file-moverservice. In thevolumessection specify the'ssd-path:/mnt/fast_storage/uploads'– the hot storage directory, from which the files will be subsequently transferred to the cold storage.Note
The
ssd-pathin the example below is a path (directory) to the SSD that will be used for the hot storage. Here and on replace thessd-pathwith an existing SSD directory.findface-multi-file-mover: ... volumes: ['./configs/findface-multi-file-mover/findface-multi-file-mover.yaml:/etc/config.yml:ro', './data/findface-multi-legacy/uploads:/mnt/slow_storage/uploads', 'ssd-path:/mnt/fast_storage/uploads']
In the
findface-multi-legacyandfindface-multi-uiservices, specify the hot storage directory'ssd-path:/var/lib/fast-storage/uploads'in thevolumessection. This should be done for the following purpose: thefindface-multi-legacyservice saves event images to the specified hot storage directory. Also, before the file moving service moves a thumbnail and a full frame event image to the cold storage, thefindface-multi-uiservice fetches them from the specified hot storage directory to display them to a user.... findface-multi-legacy: depends_on: [pgbouncer, nats, findface-sf-api, findface-counter, findface-liveness-api, etcd, findface-multi-identity-provider] environment: {ADMIN_PASSWORD: admin} image: docker.int.ntl/ntech/multi/multi/legacy:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro', './data/findface-multi-legacy/uploads:/var/lib/findface-security/uploads', 'ssd-path:/var/lib/fast-storage/uploads'] ... findface-multi-ui: depends_on: [findface-multi-legacy] image: docker.int.ntl/ntech/multi/multi/ui:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-ui/nginx-site.conf:/etc/nginx/conf.d/default.conf:ro', './data/findface-multi-legacy/uploads:/var/lib/findface-security/uploads', 'ssd-path:/var/lib/fast-storage/uploads'] ...
From the
/opt/findface-multidirectory rebuild all FindFace Multi containers.cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Start Multiple Instances of the findface-multi-file-mover Service
To provide fault tolerance, you may want to start more than one instance of the findface-multi-file-mover service. Do the following:
In the
/opt/findface-multi/docker-compose.yamlfile, specify thefindface-multi-file-moverconfiguration as many times as many service instances you plan to use, e.g.,findface-multi-file-mover: command: [python3, /app/service.py, --config=/etc/config.yml] depends_on: [findface-multi-legacy, nats-jetstream] image: docker.int.ntl/ntech/multi/multi/file-mover:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-file-mover/findface-multi-file-mover.yaml:/etc/config.yml:ro', './data/findface-multi-legacy/uploads:/mnt/slow_storage/uploads', 'ssd-path:/mnt/fast_storage/uploads'] findface-multi-file-mover2: command: [python3, /app/service.py, --config=/etc/config.yml] depends_on: [findface-multi-legacy, nats-jetstream] image: docker.int.ntl/ntech/multi/multi/file-mover:ffmulti-2.1.0 logging: {driver: journald} network_mode: service:pause restart: always volumes: ['./configs/findface-multi-file-mover/findface-multi-file-mover.yaml:/etc/config.yml:ro', './data/findface-multi-legacy/uploads:/mnt/slow_storage/uploads', 'ssd-path:/mnt/fast_storage/uploads']
From the
/opt/findface-multidirectory rebuild all FindFace Multi containers.cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Note
When saving to the hot storage is enabled, thumbnails and full frame images that were not transferred to the cold storage will not be saved by backup. If you need to save all images, copy them from the hot to the cold storage using a rsync utility. See. Back up FindFace Multi.
Migrate Data to Another Disk
High disk load may lead to delays in event arrivals. In severe cases, it might result in complete inoperability of FindFace Multi. One of the means for reducing the disk load is to migrate the FindFace Multi data storages to another disk.
In this section:
Prepare Disk
To prepare a disk for the data migration, do the following:
Create a new mount point (
/mnt/ffdatain our example).sudo mkdir /mnt/ffdata
Create a partition.
sudo parted /dev/sdb mklabel gpt mkpart primary ext4 1MiB 100% q sudo mkfs.ext4 /dev/sdb1
Learn the UUID of the partition (
sdb1in our example).sudo blkid | grep sdb1 /dev/sdb1: LABEL="data" UUID="0638ebe0-853e-43ea-8f35-bfae305695d1" TYPE="ext4" PARTUUID="8cebaacc-77d7-4757-b4c6-14147e92646c"
Add the partition to
fstabto make it automatically mount on booting.sudo vi /etc/fstab ------------ #DATA mount UUID=0638ebe0-853e-43ea-8f35-bfae305695d1 /mnt/ffdata/ ext4 auto,user,rw 0 2 -------------
Mount all the filesystems.
sudo mount -a
Migrate Photo and Video Storage
Note
We recommend that you get acquainted with FindFace Multi Data Storages before starting migration of storages.
To migrate the FindFace Multi photo and video storage, do the following:
Inside the mount point, create directories (e.g.,
/mnt/ffdata/ffuploadand/mnt/ffdata/findface-security) to store photos and recorded video chunks (if the Video Recorder was deployed). Then, move contents of the/opt/findface-multi/data/findface-upload/uploads/and the/opt/findface-multi/data/findface-multi-legacy/uploads/directories to the newly created directories.sudo mkdir /mnt/ffdata/ffupload sudo mkdir /mnt/ffdata/findface-security sudo cp -ax /opt/findface-multi/data/findface-upload/uploads/ -R /mnt/ffdata/ffupload/ sudo cp -ax /opt/findface-multi/data/findface-multi-legacy/uploads/ -R /mnt/ffdata/findface-security/ sudo rm -r /opt/findface-multi/data/findface-multi-legacy/uploads/ sudo rm -r /opt/findface-multi/data/findface-upload/uploads/
Mount the created directories (
/mnt/ffdata/ffuploadand/mnt/ffdata/findface-securityin the example) into the corresponding containers. To do so, open the/opt/findface-multi/docker-compose.yamlconfiguration file and list them in the volumes of the sections mentioned in the example instead of the default directories.sudo vi /opt/findface-multi/docker-compose.yaml findface-upload: ... volumes: ['./configs/findface-upload/40-ffupload.sh:/docker-entrypoint.d/40-ffupload.sh:ro', '/mnt/ffdata/ffupload:/var/lib/ffupload'] ... findface-multi-identity-provider: ... volumes: ['./configs/findface-multi-identity-provider/findface-multi-identity-provider.py:/etc/findface-security/config.py:ro', '/mnt/ffdata/findface-security/uploads:/var/lib/findface-security/uploads'] ... findface-multi-legacy: ... volumes: ['./configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro', '/mnt/ffdata/findface-security/uploads:/var/lib/findface-security/uploads'] ... cleaner: ... volumes: ['./configs/findface-multi-legacy/findface-multi-legacy.py:/etc/findface-security/config.py:ro', '/mnt/ffdata/findface-security/uploads:/var/lib/findface-security/uploads'] ... findface-multi-ui: ... volumes: ['./configs/findface-multi-ui/nginx-site.conf:/etc/nginx/conf.d/default.conf:ro', '/mnt/ffdata/findface-security/uploads:/var/lib/findface-security/uploads'] ... findface-multi-file-mover: ... volumes: ['./configs/findface-multi-file-mover/findface-multi-file-mover.yaml:/etc/config.yml:ro', '/mnt/ffdata/findface-security/uploads:/mnt/slow_storage/uploads']
Rebuild all FindFace Multi containers.
cd /opt/findface-multi sudo docker-compose down sudo docker-compose up -d
Migrate Main Database (PostgreSQL)
To migrate the PostgreSQL database, do the following:
Stop all FindFace Multi containers:
cd /opt/findface-multi sudo docker-compose down
Move the
/opt/findface-multi/data/postgresqldirectory to the/mnt/ffdatadirectory.sudo mv /opt/findface-multi/data/postgresql /mnt/ffdata
Specify the new path of the
postgresqldirectory (/mnt/ffdata/postgresqlin the example) to mount into thefindface-multi-postgresql-1container. To do so, open the/opt/findface-multi/docker-compose.yamlconfiguration file and list the/mnt/ffdata/postgresqldirectory in the volumes of thepostgresqlsection instead of the default/opt/findface-multi/data/postgresqldirectory.sudo vi /opt/findface-multi/docker-compose.yaml postgresql: ... volumes: ['./configs/postgresql/40-init.sql:/docker-entrypoint-initdb.d/40-init.sql:ro', '/mnt/ffdata/postgresql:/bitnami/postgresql']
Start all FindFace Multi containers.
sudo docker-compose up -d
Disable Services
You can disable the following FindFace Multi services should you no longer need them:
episodes
video archive queue manager
webhooks
Active Directory
VMS cleanup
To do so, open the /opt/findface-multi/config/findface-multi-legacy/findface-multi-legacy.py configuration file and modify the SERVICES section, setting False for the services that are no longer in use. Restart the findface-multi-findface-multi-legacy-1 container.
sudo vi /opt/findface-multi/config/findface-multi-legacy/findface-multi-legacy.py
# disable unused services to increase
# overall system performance in some cases.
SERVICES = {
"ffsecurity": {
"episodes": False,
"webhooks": False,
# use queue manager to prevent drops of video archive events
"video_archive_events_manager": False,
"active_directory": False,
"vms_cleanup": False,
}
}
sudo docker container restart findface-multi-findface-multi-legacy-1
After that, relevant sections will disappear from the web interface.
Note
A relevant section will remain if there are some entities on it (for example, webhooks on the Webhooks tab). However, the service will stop working and generating data.
Appendices
FindFace Multi Data Storages
In this section:
List of Storages
FindFace Multi uses the following data storages:
Tarantool-based feature vector database that stores object feature vectors and recognition events:
/opt/findface-multi/data/findface-tarantool-serverPostgreSQL-based main system database that stores internal system data, records, and user accounts:
/opt/findface-multi/data/postgresqlThe directory that stores photos uploaded to records, video files, full frames of events and counters, and object thumbnails:
/opt/findface-multi/data/findface-upload/uploadsThe directory that stores such event artifacts as normalized object images:
/opt/findface-multi/data/findface-multi-legacy/uploadsNote
It also stores video chunks if you are using Video Recorder.
(Only with Video Recorder) MongoDB-based database storage:
/opt/findface-multi/data/mongodb/that stores meta-information of the video chunks, including their exact location in the/opt/findface-multi/data/findface-multi-legacy/uploads.
Feature Vector Database Galleries
There are the following default galleries in the Tarantool-based feature vector database:
ffsec_body_events: feature vectors extracted from bodies detected in the video.ffsec_body_objects: feature vectors extracted from body images uploaded to the record index.ffsec_body_clusters: centroids of body clusters.ffsec_car_events: feature vectors extracted from vehicles detected in the video.ffsec_car_objects: feature vectors extracted from vehicle images uploaded to the record index.ffsec_car_clusters: centroids of vehicle clusters.ffsec_face_events: feature vectors extracted from faces detected in the video.ffsec_face_objects: feature vectors extracted from face images uploaded to the record index.ffsec_face_clusters: centroids of face clusters.ffsec_user_face: feature vectors extracted from the FindFace Multi users’ photos for face-based authentication.
User’s Guide
This chapter describes how to work with the FindFace web interface, including its advanced possibilities, and will be of interest to administrators, operators, and other users.
Getting Started
Once you have successfully deployed and configured FindFace Multi, it’s time to open the web interface, and get started. In this chapter, you can find a recommended sequence of steps that will help you harness your system’s complete functionality.
In this chapter:
Gear Up for Work
Perform the primary configuration of your system:
Log into your system.
Choose the language and dark or light theme.
Organize Cameras
Create a new camera group or use the default one. A camera group is an entity that allows you to group cameras subject to their physical location. For example, cameras at the same entrance to a building can be combined into one camera group.
Add cameras to the camera group and check their statuses.
You may also need:
Configure your system to process video from the group of cameras at their physical location. It may come in handy in a distributed architecture. Learn more.
Consider enabling event deduplication if observation scenes of cameras within the group overlap. This feature allows you to exclude coinciding object recognition events among cameras belonging to the same group. Learn more.
Organize Watch Lists and Record Index
Create a new watch list or use the default one. A watch list is an entity that allows you to classify objects (faces, bodies, vehicles) by arbitrary criteria, e.g., persona non grata, wanted, VIP, staff, etc.
Upload records and add them to the watch list either manually, in bulk via the web interface, or use the console bulk upload function.
You may also need:
Customize record content. Create additional fields, tabs, and search filters.
Start Monitoring Objects
By default, FindFace Multi is monitoring only unmatched objects. To enable a watch list monitoring, make this list active. You can also turn on sound notifications and request manual acknowledgment for the events associated with the list.
You may also need:
Enable automatic clustering of objects of the same origin: face/body images belonging to the same person, images of the same vehicle.
Support laws related to the processing of personal data of individuals (GDPR and similar). Learn more.
Organize Video Surveillance
Create a camera layout for essential video surveillance.
You may also need:
For advanced video surveillance, enable Video Recorder.
Count People and Vehicles. Measure Distance between People
Set up counters to count faces, bodies, and vehicles on connected cameras. You can also configure counters to measure distance between bodies. The Counters possibilities can apply to a wide range of situations, such as people counting in queues and waiting areas, monitoring public gatherings, crowding prevention, health protocol enforcement, traffic jam detection, and more.
FindFace Multi in Action
Automatically identify objects (faces, bodies, vehicles) in live video and check them against watch lists. Work with the event history by using various filters.
Harness the episodes. An episode is a set of identification events that feature objects of the same origin (face and body images of the same person and images of the same vehicle) detected within a certain period. As events on the Events tab show up in an arbitrary order, a large number of miscellaneous events can make the work challenging and unproductive. With the Episodes, the system uses AI to organize incoming events based on the objects similarity and detection time. This allows for the effortless processing of diverse events, even in large numbers.
Search for objects in the database of detected objects and card index. Learn more.
Search archived videos for objects under monitoring.
Manually compare two objects and verify that they match.
Build detailed reports on object recognition events, episodes, search events, clusters, counters, cameras, card index, audience, or audit logs.
Basic Maintenance
Go Further
Set up webhooks to automatically send notifications about specific events, episodes, and counter records to a given URL. In this case, when such an event occurs, FindFace Multi will send an HTTP request to the URL configured for the webhook. You can use webhooks for various purposes, for example, to notify a user about a particular event, invoke required behavior on a target website, and solve security tasks such as automated access control. Learn more.
Harness the FindFace Multi functions through HTTP API.
Web Interface Basics
Use the web interface to interact with FindFace Multi. To open the web interface, enter its basic address in the address bar of your browser, and log in.
Note
The basic address is set during deployment.
Important
To log in for the first time, use the admin account created during deployment. To create more users, refer to Role and User Management.
Tip
Take your system security up a notch with face-based authentication.
Navigation
There is a different number of items in the navigation bar depending on the user’s role.
Web Interface Language and Theme
To change the system language and a theme, click 
on a top panel.
Record Index
Record index stores records of individuals, including photos of their face and bodies, biometric data, or related documents and vehicle records, including their photo, license plate number, or related documents.
To create records in bulk, use the console bulk record upload functionality.
In this section:
Create Record
To create a record manually, do the following:
Navigate to the Record Index tab.
Click + Add new record.
Specify the record name. If necessary, add a comment.
From the Watch lists drop-down menu, select a watch list for the record (or several watch lists, one by one).
Check Record active. If a record is inactive, it is excluded from monitoring.
Save the record. You will see additional tabs appear.
On the same tab Info, attach related files.
For the individual records on the Faces tab attach images of the individual’s face and on the Bodies tab attach images of the individual’s body. Supported formats: WEBP, JPG, BMP, PNG, HEIC.
For the vehicle records on the Vehicles tab attach corresponding images.
Batch Record Upload
If there are too many records to create, you can bulk load records into the record index. Do the following:
Navigate to the Record Index tab.
Click + Add new record -> batch record upload.
Select files or folder to drop and drag to upload.
Specify the name. If necessary, add a comment.
Click Start.
Filter Records
The most frequently used filters for the record index are available in the upper part of the window.
To display the entire set of filters, click the More filters button. Here it is:
Record type: display records by given type (individuals or vehicles).
Watch lists: display records from selected watch lists.
Faces: filter records by presence of a face biometric data.
Bodies: filter records by presence of a bodies biometric data.
Vehicles: filter records by presence of a vehicle data.
License plate number: filter vehicle records by license plate number.
Filling: display only empty records, only filled or any records.
Name contains: filter records by name.
ID: display a record with a given ID.
Status: filter records by status.
You can sort out records on the list by ID.
Purge Record Index
You can purge the record index entirely or by watch lists in one click. Do the following:
Navigate Settings -> Watch Lists.
Select one or several watch lists.
Click Delete associated records.
Video Sources
To configure video-based object monitoring, add cameras to FindFace Multi, grouping them subject to their location.
Note
Privileges to create camera groups and cameras are managed in user’s permissions (see Role and User Management).
In this chapter:
Create Camera Group
Tip
A default preconfigured camera group is available in the system.
To create a group of cameras, do the following:
Navigate to the Settings tab. Click Camera Groups.
Click + Add new camera group.
On the Information tab, specify the group name. Add a comment if needed.
If you want to allocate a certain
findface-video-workerinstance to process video streams from the group, create or select one or several allocation labels.Note
To complete the allocation, list the labels in the
findface-video-worker.yamlconfiguration file. See Allocate findface-video-worker to Camera Group for details.If you want to deduplicate events from cameras that belong to the same group, i.e. exclude coinciding events, select Deduplicate events with interval and specify the deduplication interval (interval between 2 consecutive checks for event uniqueness).
Warning
Use deduplication with extreme caution. If cameras within a group observe different scenes, some objects may be skipped. See Deduplicate Events for details.
By default, video from all camera groups is processed using the generic confidence threshold. To set an individual threshold for the camera group, select Confidence Threshold and specify the threshold value.
Important
The default generic confidence threshold is optimal for the majority of recognition cases. We do not recommend changing it on your own. Be sure to consult with our technical experts prior (support@ntechlab.com).
Click Save.
On the Permissions tab, assign privileges on the camera group, specifying which user roles are allowed to change/view the camera group settings.
Click Save.
Add Camera
To add a camera, do the following:
Navigate to the Video Sources → Cameras.
Click + Add cameras or + Add.
On the Devices tab, enter the stream URL and click Next.
Enter the camera name.
Add the camera to a camera group.
Select detectors that you want to enable on this camera: faces, bodies, vehicles.
Click Add and configure. You will see additional tabs appear.
Select the camera on the list to open the processing configuration wizard. Set up the video processing parameters.
If the camera is ONVIF, select it from the list of detected devices to automatically load available settings and streams.
Video processing parameters
For each camera, you will be provided with complete statistics such as current session duration, the number of successfully posted objects, the number of objects processed with errors after the last job restart, the number of frame drops, and other data. To consult these data, click the camera and go to the Info tab.
On the General tab, you can change Camera Name, Camera group, or add Description. You can record video and enable liveness. Select detectors that you want to enable on this camera: faces, bodies, vehicles.
On the Advanced tab, fine-tune the camera:
If needed, change the video orientation.
Important
Be aware that the
findface-multi-legacyserver rotates the video using post-processing tools. It can negatively affect performance. Rotate the video via the camera functionality wherever possible.Timeout: Specify the timeout in milliseconds for posting detected objects.
Verify SSL certificate: Select to enable verification of the server SSL certificate when the object tracker posts objects to the server over https. Deselect the option if you use a self-signed certificate.
Retrieve timestamps from stream: Select to retrieve and post timestamps from the video stream. Deselect the option to post the current date and time.
Add to timestamps: Add the specified number of seconds to timestamps from the stream.
FFMPEG parameters: FFMPEG options for the video stream in the key-value format, for example, [“rtsp_transpotr=tcp”, “ss=00:20:00”].
Play speed limit: If less than zero, the speed is not limited. In other cases, the stream is read with the given
play_speed. Not applicable for live-streaming.Force input format: Pass FFMPEG format (mxg, flv, etc.) if it cannot be detected automatically.
Minimum motion intensity: Minimum motion intensity to be detected by the motion detector.
Read frames from source without drops: By default (false), if
findface-video-workerdoes not have enough resources to process all frames with objects, it drops some of them. If this option is active (true)findface-video-workerputs excessive frames on a waiting list to process them later. Set a variable to true when processing files as cameras.Process the stream once until the end or until the first error (single_pass): If true, disable restarting video processing upon error.
Router URL (router_url): IP address for posting detected objects to external video workers from
findface-video-worker. By default,'http://127.0.0.1'.
Specify the region of tracking within the camera field and region of interest (Zones). Click Save.
The region of tracking enables detecting and tracking faces/bodies and vehicles only inside a clipping rectangle. You can use this option to reduce the video object detector load. The region of interest enables posting objects detected only within its boundaries.
On the Faces, Bodies and Vehicles tabs, specify settings for each object type detector.
Size: Minimum object size in pixels to post and maximum object size in pixels to post.
Quality: The minimum quality of the face image for detection. The allowed range is from 0 to 1. The recommended reference value is 0.45, which corresponds to object images of satisfying quality. Do not change the default value without consulting with our technical experts (support@ntechlab.com).
JPEG quality: Full frame compression quality.
Full frame in PNG: Send full frames in PNG and not in JPEG as set by default. Do not enable this parameter without consulting with our technical experts (support@ntechlab.com) as it can affect the entire system functioning.
Overlap threshold: The percentage of overlap of bboxes between two serial frames so that these bboxes are considered as one track. The range of values is from 0 to 1. Do not change the default value without consulting with our technical experts (support@ntechlab.com).
Track maximum duration: The maximum approximate number of frames in a track, after which the track is forcefully completed. Enable it to forcefully terminate endless tracks, for example, tracks with faces from advertising media.
Forced termination interval: Terminate the track if no new object has been detected within the specified time (in seconds).
Send track history: Send array of bbox coordinates along with the event. May be applicable for external integrations to map the path of an object.
Crop full frame: Select to crop the full frame to the size of the ROT area before sending it for recognition. The size of the full frame will be equal to the size of ROT area.
Offline mode (overall_only): By default, the system uses the offline mode to process the video, i.e., it posts one snapshot of the best quality per track to save disk space. Disable it to receive more face snapshots if needed. If the offline mode is on, the parameters of the real-time mode are off.
Real-time mode parameters:
Note
These parameters are non-functional if the offline mode is on.
Interval: Time interval in seconds (integer or decimal) within which the object tracker picks up the best snapshot in the real-time mode.
Post first object immediately: Select to post the first object from a track immediately after it passes through the quality, size, and ROI filters, without waiting for the first
realtime_post_intervalto complete in real-time mode. Deselect the option to post the first object after the firstrealtime_post_intervalcompletes.Post track first frame: At the end of the track, the first frame of the track will be additionally sent complimentary to the overall frame of the track. May be applicable for external integrations.
Post in every interval: Select to post the best snapshot within each time interval (
realtime_post_interval) in real-time mode. Deselect the option to post the best snapshot only if its quality has improved comparing to the posted snapshot.Post track last frame: At the end of the track, the last frame of the track will be additionally sent complimentary to the overall frame of the track. May be applicable for external integrations.
Confidence threshold: A confidence threshold is used for matching the detected object with the internal entity object.
(Optional) On the GEO tab, specify the camera geographical location.
Disabled by default, the VMS cleanup settings tab allows you to configure video cleanup for a single camera. Follow this instruction Enable VMS Cleanup Settings in the UI to enable the tab in the web interface.
Custom VMS cleanup settings override the global ones for a single camera. The following custom VMS cleanup settings are available:
Clean up between tracks
Enable custom cleanup: enable custom clean up between tracks’ settings. By default, global settings are used.
Clean up between tracks: remove those intervals between tracks that do not contain recognition events.
Save tracks of events: save tracks of events for a selected object (face, body, or vehicle).
Keep events before and after: a video interval in seconds before and after a track with events that is kept along with the track.
Cleanup threshold: a minimum interval between tracks in seconds. If an interval between tracks is below the threshold, the video is not removed. A minimum allowed value for this field is 360 seconds.
Clean up between tracks from: time period (in days) during which the entire video archive is kept. After this period, the intervals between tracks will be cleaned up as configured in the settings. The value can be greater than or equal to zero.
Clean up archive
Enable custom cleanup: enable custom archive cleanup settings. By default, global settings are used.
Clean up archive: enable the entire video archive cleanup.
Archive cleanup age: a number of days the video archive is kept in the system. After that period, the video archive is removed.
Clicking on Display global settings will display global VMS cleanup settings.
See also
Upload and Process Video File
To upload and process a video, do the following:
Navigate to the Video Sources → Uploads.
Click Upload files or Upload.
Specify a URL to the online storage, or select a file. Click Upload.
Designate a camera group to which the system will attribute the object recognition events from the video. The
Video archive default Camera Groupis perfect for this task. You can also create a new camera group with basic settings specifically for this video file. Select detectors that you want to enable on this video: faces, bodies, vehicles.
Click Add and configure. The video will be uploaded and shown in the source list.
Click the video on the list to open the processing configuration wizard. Specify parameters of video processing in the same way as for cameras.
(Optional) Select a camera within that camera group to tag the object recognition events from this video more precisely.
(Optional) Configure the timestamps for object recognition events.
On the list of video uploads, click three dots → Process to start object identification. You can view object identification events on the Events and Episodes tabs by filtering the list of events by the camera group/camera associated with the video.
External Detector
A camera object can also be used for integrating an External Detector.
If necessary, change the External detector name. Specify the Initial camera group and Description, enable liveness.
Monitor Camera Operation
To monitor the operation of cameras, navigate to the Cameras tab.
Camera statuses:
Green: the video stream is being processed without errors.
Yellow: the video stream is being processed for less than 30 seconds, or one or more errors occurred when posting an object.
Red: the video stream cannot be processed.
Grey: camera disabled.
Tip
You can configure the yellow and red statuses based on the portion of dropped frames and failed object postings. To do so, modify the following parameters in the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file:
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
# max camera frames_dropped percent
'MAX_CAMERA_DROPPED_FRAMES': {'yellow': 0.1, 'red': 0.3},
# max camera objects failed percent
'MAX_CAMERA_FAILED_OBJECTS': {'yellow': 0.1, 'red': 0.3},
...
}
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker container restart findface-multi-findface-multi-legacy-1
Each created camera is associated with a so-called job, a video processing task that contains configuration settings and stream data and is assigned to findface-video-worker. This task can be restarted.
To restart a job, click Restart. In this case, the number of errors will be reset to 0.
If there are numerous cameras in the system, use the following filters:
Name contains,
Camera groups,
Activity Status (Any / Only active / Only inactive),
Status of operation (Any / Gray / Green / Read / Yellow).
Line Crossing
In this chapter:
The line crossing functionality is an integral part of the video analytics system and plays a significant role in perimeter security. It allows you to determine the violation based on the crossing of pre-defined virtual lines by people or vehicles.
In addition, the line crossing functionality allows you to collect data for further analysis of the flow of visitors in the Analytical BI service and solve the following tasks:
Perimeter control and penetration into restricted areas:
passage to prohibited areas
crossing a line in a forbidden direction
Marketing analytics – analytics of the number of people or cars crossing the line in a given direction (in a shop, in the parking lot, at a bus stop, when visiting parks and public places). For example:
analysis of the flow of visitors and determine the most popular areas inside the store
analysis of the load on the transport network depending on the day of the week and time of the day and others
Setting Up Video Sources
For working with line crossing functionality, configure cameras and uploads. Do the following:
Navigate to the Video Sources tab.
Check the camera or uploads on the list to open the processing configuration wizard.
Navigate to the Faces, Bodies or Vehicles tab, depending on the specified detectors.
Check Send track history.
Set the Track maximum duration to a value greater than 1 (the recommended value is 1000).
Click Save.
Tip
If there are a lot of objects in the video stream, and they begin to overlap each other periodically, which is why some line crossings may not be fixed, then you can increase the value of Overlap threshold in the camera settings for the desired detector.
Tip
If the objects are far enough away, and are poorly or not recognized at all, then you can reduce the value of min_size in the findface-video-worker.yaml file for the desired detector, for example, 32 or 16.
sudo vi /opt/findface-multi/configs/findface-video-worker/findface-video-worker.yaml
...
models:
cache_dir: /var/cache/findface/models_cache
detectors:
body:
fnk_path: /usr/share/findface-data/models/detector/body.jasmine_fast.018.cpu.fnk
min_size: 60
face:
fnk_path: /usr/share/findface-data/models/detector/face.jasmine_fast.003.cpu.fnk
min_size: 60
car:
fnk_path: /usr/share/findface-data/models/detector/car.gustav_normal.004.cpu.fnk
min_size: 60
...
Important
When reducing the value of min_size, the load on server resources is increased.
Create Line Crossing
To create a line crossing, do the following:
Navigate to the Line Crossing tab.
Click Create line.
On the Information tab, specify the line name. Add a comment, if needed.
Check Active.
Specify data source. On the Camera tab or Video tab, select a camera group and a camera or a video archive accordingly.
Set the intersection point of the bounding box.
On the Line tab, draw the line.
On the BI Analytics tab, set the line attributes for the Analytical BI service, if necessary.
Click Save.
You will see your list of lines on the Line Crossing tab.
By clicking the three dots, you can apply the following actions to each line crossing:
Filter events
Deactivate
Delete
Filter Line Crossing
The most frequently used filters for the line crossing are available in the upper part of the window.
To display the entire set of filters, click the More filter button. Here it is:
Name contains: filter lines by name.
ID: display a line crossing with a given ID.
Status: filter line crossing by status.
Camera group: display only line crossing from a selected group of cameras.
Cameras: display only line crossing from a selected camera.
Video archive: display only line crossing from a selected video archive.
To create a report, click Apply and create report.
Monitor Line Crossing Episodes and Events
To monitor episodes and events of line crossing, navigate to the Episodes & Events tab. Click Episodes or Events.
You can navigate directly to the Events tab by clicking three dots → Filter face events/Filter body events/Filter vehicle events on the list of lines.
An episode ticket contains information about line crossings that have been created in all of its events (AB or BA).
Also for an event or episode, a red crossing indicates a reverse crossing: when the object has crossed the same line an even number of times and the last crossing is in reverse of the direction from which the first crossing began.
On the full frame of an episode with a line crossing, the best event of an episode with the bbox of an object is displayed, and the line with the track is displayed from the event that had a crossing. The blue color shows the beginning of the track, and the red color shows the end of the track. If there are several events with crossing in the episode, the line and track will be displayed from the first event with crossing in the episode.
Events and Episodes of Object Recognition
To monitor the real-time object identification in live videos, navigate to the Episodes & Events and use the Episodes and Events tabs. Besides monitoring, both tabs allow you to access the history of identification events.
Tip
To perform the object identification in archived videos, see Upload and Process Video File.
Work with Events
This section is about the Events tab.
Important
You can enable sound notifications for events related to specific watch lists. In some browsers, the tab with events has to remain in focus to get a sound played. To put a tab in focus, open it, and click anywhere on the page.
In this chapter:
View Events
Once an object is detected, you will see a notification on one of the event lists: Faces, Bodies, or Vehicles, subject to the object type.
A notification can feature different pieces of information, depending on whether a detected object has a match in the record index:
Match not found: the normalized object image, detection date and time, camera group name, watch lists.
Match found: the normalized object image, reference object photo from the related record, name, the similarity between matched objects, comment from the record, watch lists, detection date and time, camera group, attributes.
Important
In order to pause the notification thread, click 
above the list of events.
When working with events, the following filters may come in handy:
Note
Some filters from the list below may be hidden, subject to enabled recognition features.
Object: display events only for faces, bodies or vehicles.
Matches: display events only with/without matches, or all events.
Watch lists: display events only for a selected watch list.
Acknowledged: display only acknowledged/unacknowledged events, or all events.
Camera groups: display only events from a selected group of cameras.
Cameras: display only events from a selected camera.
Date and time: display only events that occurred within a certain period.
Video Archive ID: display events from the video archive with a given ID.
Record name: display events with a given record name.
Episode ID: display events from the episode with a given ID.
Event ID: display an event with a given ID.
Event’s best shot: display all events of a track, only events with real-time snapshots, only one event with the best snapshot at the end of a track.
Line: display only events from a selected line.
Line crossing direction: filter events by line crossing direction, none, A → B, B → A.
Has reverse crossing: filter events by having/not having a reverse crossing.
Specific filters for faces
Age: display events with people of a given age.
Beard: filter events by the fact of having a beard.
Emotions: display events with given emotions.
Gender: display events with people of a given gender.
Glasses: filter events by the fact of wearing glasses.
Liveness: filter events by face liveness.
Face mask: filter events by the fact of wearing a face mask.
Head turn: filter events by degree of head turn.
Head tilt: filter events by degree of head tilt.
Specific filters for bodies
Gender by body: display only events with people of a given gender or all events.
Age by body: display only events with people of a given age.
Headwear: display only events with a person wearing headgear of a given type: hat/cap, hood/headscarf, none.
Vest: display only events with a person wearing a vest of a given color.
Helmet: display only events with a person wearing a helmet of a given color.
Upper clothes color: display only events with a person wearing a top of a given color.
Lower clothes color: display only events with a person wearing a bottom of a given color.
Upper clothes type: display only events with a person wearing upper body wear of a given specific type: jacket, coat, sleeveless vest, sweatshirt, T-shirt, shirt, dress.
Lower body clothes: display only events with a person wearing lower body wear of a given type: pants, skirt, shorts, obscured.
Upper body clothes: display only events with a person wearing upper body wear of a given generalized category: long sleeves, short sleeves, no sleeve.
Bag on the back: display only events with a person wearing/not wearing a bag on the back.
Bag in hand: display only events with a person wearing/not wearing a bag in hand.
Specific filters for vehicles
Make: filter vehicle events by vehicle make.
Model: filter vehicle events by vehicle model.
Vehicle body type: display only events with vehicles of a given body type: minivan, limousine.
Vehicle body color: display only events with vehicles of a given color.
Country: display only events with vehicles registered in a given country.
License plate number: display an event with a given plate number.
Region: display only events with vehicles registered in a given region.
License plate color: display only events with a given license plate color.
Special vehicle: display only events with vehicles belonging to a given type: police, fire service and EMERCOM vehicles, gas rescue and emergency services, military, municipal vehicles, and others.
Vehicle category: display only events with vehicles belonging to a given category: motorcycle, scooter, car, car with a trailer, truck, truck with a trailer, bus, articulated bus, and others.
Vehicle weight and body size: display only events with vehicles of a given weight and body size.
Vehicle orientation: display only events with vehicles of a given orientation.
Note
License plate region and color are predicted for the United Arab Emirates (UAE) only. The values of these attributes will be marked as unknown for other countries.
Event Ticket. Acknowledging Event
In order to navigate to an event ticket from the list of events, click on the recognition result in a notification.
An event ticket contains the same data as a relevant notification. It also allows for acknowledging the event. To do so, click Acknowledge to change the event acknowledgment status to Acknowledged.
Tip
If a detected object has a match in the record index, you can navigate to the relevant record by clicking on the name in the event ticket.
Tip
In order to acknowledge all events, click Acknowledge all above the list of events.
Note
Event acknowledgment can be automated for selected watch lists.
Event Ticket. Object Search
FindFace Multi allows you to search detected objects through the list of events. To navigate from an event ticket to the search tab, click Search.
See also
Create or Enrich a Record from the Event
You can either enrich a record with a photo of the detected object or create a new record from the event.
In the detailed event view click on Use for record. If a detected object has a match in the record index, you will be able to either create a new record from the event or add a photo of a detected object to an existing record.
For an unmatched object there’s only an option to create a new record from the event.
Work Time Reports on Events
See Reports.
Organize Events with Episodes
This section is about the Episodes tab.
See also
An episode is a set of identification events that feature objects of the same origin (face and body images of the same person and images of the same vehicle) detected within a specific period. As events on the Events tab show up in an arbitrary order, a large number of miscellaneous events can make the work difficult and unproductive. With the episodes, the system uses AI to organize incoming events based on the objects’ similarity and detection time. This allows for easy processing of diverse events, even in large numbers.
In this chapter:
About Episodes
There are two stages of an episode lifecycle:
LIVE: an episode is currently active, to which new events can be possibly added.
Closed: an episode is closed, no events can be added.
Episodes are classified as individual episodes and vehicle episodes. Individual episodes feature face and body images of people, detected within a specific period, while vehicle episodes contain vehicle images and license plate.
Grant Rights for Episodes
A user receives a notification of a new episode if they have rights to the first event. Viewing new events in the episode also requires proper rights.
The right to an event consists of the rights for a corresponding camera and watch list.
Note
To see unmatched events, you only need the rights to a camera.
To manage the rights of a role for the entire Episode entity, open permissions for this role and adjust the following permissions:
humanepisode: individuals episodescarepisode: vehicle episodes
Tip
View Episodes
You can find the list of episodes with filters and statistics on the Episodes & Events -> Episodes tab. Episodes are sorted subject to the object’s origin: there are individual episodes and vehicle episodes.
Once an object is detected, it is either added to an existing LIVE episode or used as a starting point for a new episode. Each episode is assigned an identifier which can be later used to filter events and episodes.
When working with the episodes, the following default filters may come in handy:
Note
Some filters from the list below may be hidden, subject to enabled recognition features.
Object: display events only for individuals or vehicles.
Matches: faces: display episodes only with/without face matches, or all episodes.
Matches: bodies: display episodes only with/without body matches, or all episodes.
Matches: display episodes only with/without vehicle matches, or all episodes.
Watch lists: display episodes only for a selected watch list.
Acknowledged: display only acknowledged/unacknowledged episodes, or all episodes.
Camera groups: display only episodes from a selected group of cameras.
Cameras: display only episodes from a selected camera.
Date and time: display only episodes that occurred within a certain period.
Video Archive ID: display episodes related to the video archive with a given ID.
Record name: display only episodes with a given record name.
Episode ID: display an episode with a given ID.
Count events: display only episodes with a given number of events.
Specific filters for faces
Age: display episodes with people of a given age.
Beard: filter episodes by the fact of having a beard.
Emotions: display episodes with given emotions.
Gender: display episodes with people of a given gender.
Glasses: filter episodes by the fact of wearing glasses.
Liveness: filter episodes by face liveness.
Face mask: filter episodes by the fact of wearing a face mask.
Specific filters for bodies
Gender by body: display episodes with a body of a given gender.
Age by body: display episodes with a body of a given age.
Headwear: display only episodes with a person wearing headgear of a given type: hat/cap, hood/scarf, none.
Vest: display only episodes with a person wearing a vest of a given color.
Vest score: display only episodes with a person wearing a vest within a given score.
Helmet: display only episodes with a person wearing a helmet of a given color.
Helmet score: display only episodes with a person wearing a helmet within a given score.
Upper clothes color: display only episodes with a person wearing a top of a given color.
Lower clothes color: display only episodes with a person wearing a bottom of a given color.
Upper clothes type: display only episodes with a person wearing upper body wear of a given specific type: jacket, coat, sleeveless, sweatshirt, T-shirt, shirt, dress.
Lower body clothes: display only episodes with a person wearing lower body wear of a given type: pants, nondescript, skirt, shorts.
Upper body clothes: display only episodes with a person wearing upper body wear of a given generalized category: long sleeves, short sleeves, no sleeve.
Bag on the back: display only episodes with a person wearing/not wearing a bag on the back.
Bag in hand: display only episodes with a person wearing/not wearing a bag in hand.
Specific filters for vehicle episodes
Make: filter vehicle episodes by vehicle make.
Model: filter vehicle episodes by vehicle model.
Vehicle body type: display only episodes with vehicles of a given body type: minivan, limousine.
Vehicle body color: display only episodes with vehicles of a given color.
Country: display only episodes with vehicles registered in a given country.
License plate number: display an episode with a given plate number.
Region: display only episodes with vehicles registered in a given region.
License plate color: display only episodes with a given license plate color.
Special vehicle: display only episodes with vehicles belonging to a given type: taxi, public road transport, carsharing, ambulance, police, fire service and emergencies ministry vehicles, gas rescue and emergency services, military, municipal vehicles, other or not special.
Vehicle category: display only episodes with vehicles belonging to a given category: motorcycle, scooter, car, car with a trailer, truck, truck with a trailer, bus, articulated bus, other.
Vehicle weight and body size: display only episodes with vehicles of a given weight and body size.
Note
License plate region and color are predicted for the United Arab Emirates (UAE) only. The values of these attributes will be marked as unknown for other countries.
To view episode events, click the episode on the list. You will be redirected to the Matched Episode or Unmatched Episode tab with the episode information and related events.
Work with the Events tab as described in Work with Events.
Event and Episode Acknowledging
To acknowledge an entire episode, click Acknowledge for this episode or 
on the list. As a result, all events in the episode will be automatically acknowledged, including those that are yet-to-appear (in the case of a LIVE episode).
An episode is also automatically acknowledged after acknowledging all its events one by one.
Episode Settings
To configure the episodes, use the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file. Find the following parameters in the FFSECURITY section:
FACE_EPISODES_THRESHOLD,BODY_EPISODES_THRESHOLD,CAR_EPISODES_THRESHOLD: Similarity thresholds that the system are using when searching for recent events in order to construct an episode. The default values are set to the optimum. If necessary, you can change them. Be sure to consult with our technical experts prior (support@ntechlab.com).FACE_EPISODE_SEARCH_INTERVAL,BODY_EPISODE_SEARCH_INTERVAL,CAR_EPISODE_SEARCH_INTERVAL: The period preceding an event, for which the system searches the feature vector database for already existing events with similar objects. If no such event is found, the system creates a new episode. Otherwise, it sorts out the 100 most recent similar objects and picks up the most relevant event from a LIVE episode.EPISODE_MAX_DURATION: The maximum episode duration in seconds. After this time, an episode automatically closes.EPISODE_EVENT_TIMEOUT: The maximum time in seconds since the last event has been added to an episode. After this time, an episode automatically closes.EPISODE_KEEP_ONLY_BEST_EVENT: When closing an episode, delete all events in it, except the one with the best object. Use this option to save disk space.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
...
FFSECURITY = {
...
# episodes specific matching threshold that is used to join faces in an episode
'FACE_EPISODES_THRESHOLD': 0.673, # FAR = 1.00E-06 # model: [mango_320]
'BODY_EPISODES_THRESHOLD': 0.75, # model: [durga]
'CAR_EPISODES_THRESHOLD': 0.61, # model: [bottas]
...
# when closing episode, delete all events except the best episode event
'EPISODE_KEEP_ONLY_BEST_EVENT': False,
# delete episode events after delay in seconds
'EPISODE_DELETE_EVENTS_DELAY': 60,
...
# maximum event age in seconds than could be added to an episode.
'FACE_EPISODE_SEARCH_INTERVAL': 60,
'BODY_EPISODE_SEARCH_INTERVAL': 60,
'CAR_EPISODE_SEARCH_INTERVAL': 60,
# maximum episode duration (episode is closed after)
'EPISODE_MAX_DURATION': 300,
# if no new event added to an episode during this timeout, episode will be closed.
'EPISODE_EVENT_TIMEOUT': 30,
...
}
...
Restart the findface-multi-findface-multi-legacy-1 container.
sudo docker container restart findface-multi-findface-multi-legacy-1
Search Objects in System
FindFace Multi allows you to search for individuals throughout the entire system.
To find an individual, do the following:
Navigate to the Search tab.
Specify an object to search for in one of the following ways:
by event’s URL or ID
by record’s URL or ID
by cluster’s URL or ID
by uploading a photo
If you specified a record URL, select a photo from it. If there are multiple photos, you can select some or all of them. Click the Apply button. When you select several objects from the record index, the search range will narrow. The search results will only contain the objects similar to those on the selected photos and within the preset search threshold.
If you uploaded a photo, it will be displayed in the new window. If there are multiple objects in the image, select the one of your interest. Click Search.
You will see the search results appear. If necessary, you can narrow down your search by specifying a watch list, similarity threshold, etc.
Verify Two Objects
FindFace Multi allows you to compare two objects and verify that they match. Do the following:
Navigate to the Verify tab.
Specify two objects to verify in one of the following ways:
by event’s URL
by uploading a photo
If there are multiple objects in the image, select the one of your interest.
You will see the probability that the objects match.
Face, Body, Vehicle Counters. Distance Measurement
Important
To be able to count human bodies (silhouettes) or vehicles, you first have to enable body detection or vehicle detection.
FindFace Multi allows you to count faces, bodies and vehicles on connected cameras. This functionality can apply to a wide range of situations, such as people counting in queues and waiting areas, monitoring public gatherings, crowding prevention, traffic jam detection, and more.
The counting method is based on time slices, which means that the system counts faces, bodies and vehicles in static screenshots taken with a given count interval.
You can opt for counting faces/bodies/vehicles either on each camera independently, or collectively on all selected cameras.
In this section:
Create Counter
To set up a counter, do the following:
Navigate to the Counters tab.
Click + Add new counter.
Specify the counter name.
Select one or several camera groups for counting.
Select one or several cameras for counting.
Check Faces to count faces.
Check Bodies to count bodies. Body detection has to be enabled.
Check Vehicles to count vehicles. Vehicle detection has to be enabled.
Specify the interval in seconds between two consecutive screenshots used for counting.
Click Save.
Check Distance measurement to count the distance between bodies. This option becomes available only if Bodies detector is checked.
Make sure that the counter is activated.
Click Save.
(Optional) Navigate to the ROI tab to specify the face/body/vehicle tracking region within the camera(s) field. Select a camera and specify a region of interest. Click Save.
Calibrate Counter for Distance Measurement
If Distance measurement is enabled, you should perform the counter calibration.
For counter calibration:
Ask a person to stand before the selected camera at full height.
Ask a person to walk through the area visible by the camera, along which the distances will be calculated. To achieve the best quality, the person must walk on a flat surface.
Enter the counter calibration timeframe (from 15s to 300s) and click Calibrate counter.
If you need to calibrate the counter for several cameras, perform steps 1-3 for each camera selected in the Cameras field.
If you are unsatisfied by the result, click Delete calibration and perform the calibration once more.
Monitor Counter Operation
To monitor the operation of counters, navigate to the Counters tab.
Counter statuses:
Green: the counter is running without errors, all frames from all cameras of the counter have been received for a certain period.
Yellow: the number of errors is less than 70%.
Red: the number of errors is over 70%.
Grey: the counter is disabled.
You may use the counter status filter to find counters with technical problems, for example, if the counter could not receive frames due to loss of connection with the camera.
By selecting a counter, you can filter the list of frames and leave only frames with errors. The error tooltip is displayed when pointing over the frame id.
Tip
You can configure the yellow and red statuses based on the portion of failed counter records. To do so, modify the MAX_COUNTER_ERROR_RECORDS parameter in the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file:
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
FFSECURITY = {
...
# Counter health status config:
# max percent of camera records with errors
'MAX_COUNTER_ERROR_RECORDS': {'yellow': 0.3, 'red': 0.5},
...
}
Work with Counter Records
Static screenshots taken by a counter, with the number of faces, bodies and vehicles in them, are saved as counter entries. If you have enabled the distance measurement, each record will also contain a minimum, average, and maximum detected distance in meters.
If the counter is running with errors, the system will be creating blank records with an error message.
To see the counter records, navigate to the Counters tab. Click on the counter. Navigate to the Screenshots tab.
You can scale the screenshots by clicking on them. If distance measurement is enabled, the screenshot will contain boxes around bodies and the distance between them.
To work with the counter list, use the following filters:
Name contains: filter counters by name.
Counter ID: display a counter with a given ID.
Camera groups: display counters from selected camera groups.
Cameras: display counters from a selected camera.
Distance: filter counters with/without distance measurement or any.
Status: filter counters by status.
To work with counter screenshots, use the following filters:
Cameras: display screenshots from a selected camera.
Date and time: display screenshots that occurred within a certain period.
Error: filter screenshots with/without errors.
Count faces: display screenshots with the number of faces within a certain interval.
Count bodies: display screenshots with the number of bodies within a certain interval.
Count vehicles: display screenshots with the number of vehicles within a certain interval.
Counter entry ID: display screenshots with a given counter entry ID.
Minimum distance: display screenshots with a minimum distance within a certain interval.
Maximum distance: display screenshots with a maximum distance within a certain interval.
Average distance: display screenshots with an average distance within a certain interval.
Set Webhook for Counter
To take it up a notch, configure a webhook for counter records with a specific number of faces, bodies and vehicles.
Configure Counters
To configure counters, open the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file and modify the following parameters:
COUNTERS_DEDUP_OPTIONS: a set of options that help avoid object duplication in multi-camera counters, i.e., the situation when the same object is counted on several cameras at the same time, leading to the overly increased counting result. Two options are available for each object type (face,body,car):enabled- enables object deduplication,threshold- defines the minimum level of similarity between objects for the system to consider them duplicates.COUNTERS_SAVE_FULLFRAMEdetermines saving options of full frames in counters:always,detect- only save if faces, bodies or vehicles have been detected,never.COUNTERS_FULLFRAME_JPEG_QUALITY: JPEG quality of full frames.COUNTERS_THUMBNAIL_JPEG_QUALITY: JPEG quality of thumbnails.COUNTERS_ROI_INTERSECTION_THRESHOLD: required percentage of the rectangle around the object (a.k.a.bbox) intersection with the region of interest defined for the counter.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
# deduplicate objects on multi-camera counters
'COUNTERS_DEDUP_OPTIONS': {
'face': {
'enabled': False,
'threshold': 0.714, # model: [mango_320]
},
'body': {
'enabled': False,
'threshold': 0.65, # model: [durga]
},
'car': {
'enabled': False,
'threshold': 0.65, # model: [bottas]
},
},
# counters full frame saving options:
# `always` - save always
# `detect` - save only if faces or bodies have been detected
# `never` - never save full frames
'COUNTERS_SAVE_FULLFRAME': 'always',
'COUNTERS_FULLFRAME_JPEG_QUALITY': 75,
'COUNTERS_THUMBNAIL_JPEG_QUALITY': 75,
# required percentage of bbox intersection with ROI
'COUNTERS_ROI_INTERSECTION_THRESHOLD': 0.75,
...
Be sure to restart the findface-multi-findface-multi-legacy-1 container after making changes.
sudo docker container restart findface-multi-findface-multi-legacy-1
Face, Body, Vehicle Clusters
FindFace Multi supports automatic clustering of objects of the same origin:
Face images belonging to the same person form a face cluster.
Body images belonging to the same person form a body cluster.
Images of the same vehicle constitute a vehicle cluster.
Aggregated cluster galleries of faces, bodies, and vehicles are available on the Clusters tab.
Note
If a face cluster or a body cluster matches a person record, such a cluster will automatically appear in that record. Similarly, a vehicle cluster will be saved to the corresponding vehicle record.
Important
By default, the object clustering is disabled. Enable and configure it via the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file.
In this section:
Clustering Algorithms
FindFace Multi uses the following object clustering algorithms:
Real-time clustering. This clustering algorithm processes episodes to select suitable object images and clusters the selected ones. It works on the fly after an episode is closed. The real-time clustering results are dynamically displayed on the Clusters tab and in the relevant record.
Not all episodes are used for clustering. If an episode meets all requirements (see the details below), the system forms a cluster in the following way:
Selects the best quality event.
Creates a new entity
cluster eventin the main system database PostgreSQL. The entity contains the selected event metadata, a link to the parent episode, an object feature vector, and a thumbnail.Searches for a matching object centroid in the
cluster_eventsgallery of the Tarantool feature vector database. An object centroid is a virtual feature vector averaged across all twin objects that have been detected so far (for example, a face centroid is a feature vector averaged across all face images of the same person). The system updates a matching centroid using the new event if such a centroid was found or creates a new centroid otherwise.
Scheduled clustering. This clustering algorithm works over and revises the cluster events created during the real-time clustering. This algorithm improves the cluster centroid quality as the centroid will be averaged across a more extensive array of accumulated feature vectors. The results of the scheduled clustering are displayed on the Clusters tab and in the relevant record after each scheduled iteration.
Use the RRULE format to define the schedule in the
CLUSTERS_CLUSTERIZATION_SCHEDULEparameter at/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py. Late-night hours are preferred as the scheduled clustering takes up a lot of CPU resources and time.Important
The scheduled clustering completely overwrites the cluster galleries, including ids, unless you pin specific clusters by enabling the
CLUSTERS_AUTO_PIN_HEURISTICSandPIN_MATCHED_CLUSTERSparameters (see below).
Enable and Configure Clustering
By default, the clustering is disabled. To enable and configure it, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file. Find theClusters configurationsection.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py # -- Clusters configuration -- 'ENABLE_NIGHT_CLUSTERIZATION': False, 'ENABLE_REALTIME_CLUSTERIZATION': True, # rrule (recurrence rule) for scheduling "night" clusterization 'CLUSTERIZATION_SCHEDULE': 'RRULE:FREQ=DAILY;INTERVAL=1;WKST=MO;BYHOUR=0;BYMINUTE=0', # clusterize only selected objects types (for realtime and nightly clusterization) # available are: face, body, car 'CLUSTERIZE_OBJECT_TYPES': ['face'], # keep no more than N the best quality events in centroid (None to disable) 'CENTROID_MAX_SIZE': None, # save cluster events without emben and images (only keep thumbnail for the best event) 'LONG_LIVING_CLUSTER_EVENTS': False, # skip clusterization if unpinned cluster events count is greater than this value 'CLUSTERIZATION_MAX_CLUSTER_EVENTS': None, # create cluster only from cluster events in current case or only from cluster events without case 'ISOLATE_CASE_CLUSTERS': True, # cluster event to cluster matching confidence threshold 'FACE_CLUSTER_CONFIDENCE_THRESHOLD': 0.714, # model: [mango_320] 'BODY_CLUSTER_CONFIDENCE_THRESHOLD': 0.65, # model: [durga] # minimum required event quality for cluster creation 'FACE_CLUSTER_EVENT_MIN_QUALITY': 0.5, # model: [quality_fast.v1] 'BODY_CLUSTER_EVENT_MIN_QUALITY': 0.6, # model: [pedattr.quality.v0] 'CAR_CLUSTER_EVENT_MIN_QUALITY': 0.73, # model: [carattr.quality.v0] # discard cluster event if `max_centroids` similar centroids found with confidence greater than `confidence` 'FACE_CLUSTER_MAX_N_SIMILAR': {'enabled': False, 'max_centroids': 5, 'confidence': 0.714}, 'BODY_CLUSTER_MAX_N_SIMILAR': {'enabled': False, 'max_centroids': 5, 'confidence': 0.65}, # minimum required object size in pixels for cluster creation 'FACE_CLUSTER_EVENT_MIN_SIZE': 50, 'BODY_CLUSTER_EVENT_MIN_SIZE': 50, 'CAR_CLUSTER_EVENT_MIN_SIZE': 50, # minimum required number events in episode for cluster creation 'FACE_CLUSTER_EVENT_MIN_EPISODE_EVENTS': 1, 'BODY_CLUSTER_EVENT_MIN_EPISODE_EVENTS': 1, 'CAR_CLUSTER_EVENT_MIN_EPISODE_EVENTS': 1, # age feature threshold for cluster creation 'FACE_CLUSTER_EVENT_MIN_AGE_THRESHOLD': 16, .... # pinned clusters keep their id and events after reclusterization 'CLUSTERS_AUTO_PIN_HEURISTICS': { 'face': { # pin clusters with `value` minimum cluster events 'min_events': {'enabled': True, 'value': 10}, # cluster's centroid similarity confidence is less than 'max_centroid_similarity_threshold': {'enabled': True, 'value': 0.54}, # mango_320 # minimum average event's quality 'min_average_events_quality': {'enabled': True, 'value': 0.45}, }, 'body': {}, 'car': {}, }, # always pin clusters with matched events (not affected by heuristics above) 'PIN_MATCHED_CLUSTERS': False, ...
Enable the real-time clustering by setting
ENABLE_REALTIME_CLUSTERIZATION: True.If necessary, enable the scheduled clustering by setting
ENABLE_NIGHT_CLUSTERIZATION: True.Important
Enabling the scheduled clustering only makes sense if the real-time clustering is enabled. Otherwise, the system won’t form any new clusters, since only the real-time clustering produces unique cluster events.
... # -- Clusters configuration -- 'ENABLE_NIGHT_CLUSTERIZATION': True, 'ENABLE_REALTIME_CLUSTERIZATION': True, ...
If necessary, specify a recurrence rule (RRULE) for the scheduled clustering. If the RRULE is not specified, the clustering automatically starts at 00:00 GMT.
Tip
See the RRULE calculator here.
# rrule (recurrence rule) for scheduling clusters clusterization 'CLUSTERIZATION_SCHEDULE': 'RRULE:FREQ=DAILY;INTERVAL=1;WKST=MO;BYHOUR=0;BYMINUTE=0',
By default, the system forms only face clusters. To enable cluster formation of bodies and vehicles, add relevant object types to the following line:
# available are: face, body, car 'CLUSTERIZE_OBJECT_TYPES': ['face','body','car'],
If necessary, modify the minimum number of events in the episodes used for clustering. It’s 1 by default. Do so separately for each object type.
# minimum required number events in episode for cluster creation 'FACE_CLUSTER_EVENT_MIN_EPISODE_EVENTS': 3, 'BODY_CLUSTER_EVENT_MIN_EPISODE_EVENTS': 3, 'CAR_CLUSTER_EVENT_MIN_EPISODE_EVENTS': 2,
If necessary, modify the minimum quality of the object images used for clustering. Do so separately for each object type.
Note
As this setting requires a high level of expertise and knowledge, we highly recommend consulting with our technical experts prior.
# minimum required event quality for cluster creation 'FACE_CLUSTER_EVENT_MIN_QUALITY': 0.5, # model: [quality_fast.v1] 'BODY_CLUSTER_EVENT_MIN_QUALITY': 0.6, # model: [pedattr.quality.v0] 'CAR_CLUSTER_EVENT_MIN_QUALITY': 0.73, # model: [carattr.quality.v0]
If necessary, modify the confidence threshold for matching a cluster event and a cluster.
Warning
Be sure to consult with our experts by support@ntechlab.com before altering this parameter.
# cluster event to cluster matching confidence threshold 'FACE_CLUSTER_CONFIDENCE_THRESHOLD': 0.714, # model: [mango_320] 'BODY_CLUSTER_CONFIDENCE_THRESHOLD': 0.65, # model: [durga]
The scheduled clustering completely overwrites all created clusters. You can “pin” specific clusters, i.e. keep them and associated cluster events, including their IDs, intact. To do so, use the following settings:
Note
These settings are independent. Apply both if necessary.
Note
These settings do not affect the real-time clustering. It will continue to create new cluster events for the pinned clusters.
CLUSTERS_AUTO_PIN_HEURISTICS: setTrueorFalsefor the following options and specify the corresponding values:Note
Do so for each object type if applicable.
min_events: pin a cluster when the number of associated cluster events exceeds the given minimum value.max_centroid_similarity_threshold: pin a cluster if the similarity between its centroid and centroids of other clusters is less than the given threshold. If a cluster looks similar to some other clusters, chances are these clusters belong to the exact person/vehicle. In this case, the system won’t pin such a cluster to have an opportunity to re-cluster it. On the contrary, dissimilar clusters will be pinned.min_average_events_quality: pin a cluster if the average quality of associated cluster events is greater than the given minimum value.
# pinned clusters keep their id and events after reclusterization 'CLUSTERS_AUTO_PIN_HEURISTICS': { 'face': { # pin clusters with `value` minimum cluster events 'min_events': {'enabled': True, 'value': 10}, # cluster's centroid similarity confidence is less then 'max_centroid_similarity_threshold': {'enabled': True, 'value': 0.54}, # minimum average event's quality 'min_average_events_quality': {'enabled': True, 'value': 0.45}, }, 'body': {}, 'car': {}, },
Enable the
PIN_MATCHED_CLUSTERSparameter to pin matched clusters and the associated cluster events.# always pin clusters with matched events (not affected by heuristics above) 'PIN_MATCHED_CLUSTERS': True,
If necessary, specify the maximum number of cluster events in the clusters that remain non-pinned. After this number is reached, the scheduled clustering will be automatically disabled.
# skip clusterization if unpinned cluster events count is greater than this value 'CLUSTERIZATION_MAX_CLUSTER_EVENTS': None,
Restart the
findface-multi-findface-multi-legacy-1container. You will see the Clusters tab appear in the FindFace Multi web interface.sudo docker container restart findface-multi-findface-multi-legacy-1
Work with Cluster Galleries
View and Filter Clusters
To view the cluster galleries, navigate to the Clusters tab.
When working with the cluster galleries, the following filters may come in handy:
Note
Some filters from the list below may be hidden, depending on which recognition features are enabled.
Object: display clusters only for faces, bodies or vehicles.
Matches: display clusters only with/without matches or any.
Watch lists: display only clusters for a selected watch list.
Camera groups: display only clusters from a selected group of cameras.
Cameras: display only clusters from a selected camera.
Record name: display only clusters with a given name.
Date and time: display only clusters formed within a certain period.
First cluster event: display only the first cluster event within a certain period.
Cluster event: display only cluster event within a certain period.
ID: display a cluster with a given ID.
Specific filters for face clusters
Age: display clusters with people of a given age.
Beard: filter clusters by the fact of having a beard.
Emotions: display clusters with given emotions.
Gender: display clusters with people of a given gender.
Glasses: filter clusters by the fact of wearing glasses.
Liveness: filter clusters by face liveness.
Face mask: filter clusters by the fact of wearing a face mask.
Specific filters for body clusters
Gender by body: display only events with people of a given gender or all events.
Age by body: display only events with people of a given age.
Headwear: display only events with a person wearing headgear of a given type: hat/cap, hood/scarf, none.
Vest: display only events with a person wearing a vest of a given color.
Vest score: display only events with a person wearing a vest within a given score.
Helmet: display only events with a person wearing a helmet of a given color.
Helmet score: display only events with a person wearing a helmet within a given score.
Upper clothes color: display only events with a person wearing a top of a given color.
Lower clothes color: display only events with a person wearing a bottom of a given color.
Upper clothes type: display only events with a person wearing upper body wear of a given specific type: jacket, coat, sleeveless, sweatshirt, T-shirt, shirt, dress.
Lower body clothes: display only events with a person wearing lower body wear of a given type: pants, nondescript, skirt, shorts.
Upper body clothes: display only events with a person wearing upper body wear of a given generalized category: long sleeves, short sleeves, no sleeve.
Specific filters for vehicle clusters
Make: filter vehicle events by vehicle make.
Model: filter vehicle events by vehicle model.
Vehicle body type: display only events with vehicles of a given body type: minivan, limousine.
Vehicle body color: display only events with vehicles of a given color.
Country: display only events with vehicles registered in a given country.
License plate number: display an event with a given plate number.
Region: display only events with vehicles registered in a given region.
License plate color: display only events with a given license plate color.
Special vehicle: display only events with vehicles belonging to a given type: police, fire service and emergencies ministry vehicles, gas rescue and emergency services, military, municipal vehicles, other.
Vehicle category: display only events with vehicles belonging to a given category: unknown, motorcycle, scooter, car, car with a trailer, truck, truck with a trailer, bus, articulated bus, other.
Vehicle weight and body size: display only events with vehicles of a given weight and body size.
Note
License plate region and color are predicted for the United Arab Emirates (UAE) only. The values of these attributes will be marked as unknown for other countries.
Click a cluster to see the associated cluster events. You will be redirected to the Cluster events page.
Merge and Delete Clusters
To manually merge several clusters, select them one by one and click Merge.
Note
By default, vehicle clusters are formed only by license plate. In this case, it is impossible to merge two vehicle clusters with different license plates.
To delete a cluster, select it and click Delete.
Make Cluster Gallery Static
Sometimes it’s necessary to finalize the object clustering at a certain point in time and then operate with a static gallery of formed clusters.
To display the Clusters tab upon the disabled clustering, do the following:
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyconfiguration file.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
Add the line
"clusters": Trueto theSERVICESsection, as shown in the example below:... SERVICES = { "ffsecurity": { ... "clusters": True, } ...
Disable the real-time and scheduled clustering processes.
... 'ENABLE_NIGHT_CLUSTERIZATION': False, 'ENABLE_REALTIME_CLUSTERIZATION': False,
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Manual Clustering
To manually launch the clustering process, use the run_clusterization utility.
You can invoke the run_clusterization help message by executing:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py run_clusterization --help
usage: manage.py run_clusterization [-h]
[--object-types OBJECT_TYPES [OBJECT_TYPES ...]]
[--force] [--configuration CONFIGURATION]
[--version] [-v {0,1,2,3}]
[--settings SETTINGS]
[--pythonpath PYTHONPATH] [--traceback]
[--no-color] [--force-color]
[--skip-checks]
optional arguments:
-h, --help show this help message and exit
--object-types OBJECT_TYPES [OBJECT_TYPES ...]
Clusterize selected object types. Uses
CLUSTERIZE_OBJECT_TYPES from config if not provided.
Allowed types: face, body, car
--force Force clusterization even if
CLUSTERIZATION_MAX_CLUSTER_EVENTS condition is met
--configuration CONFIGURATION
The name of the configuration class to load, e.g.
"Development". If this isn't provided, the
DJANGO_CONFIGURATION environment variable will be
used.
--version show program's version number and exit
-v {0,1,2,3}, --verbosity {0,1,2,3}
Verbosity level; 0=minimal output, 1=normal output,
2=verbose output, 3=very verbose output
--settings SETTINGS The Python path to a settings module, e.g.
"myproject.settings.main". If this isn't provided, the
DJANGO_SETTINGS_MODULE environment variable will be
used.
--pythonpath PYTHONPATH
A directory to add to the Python path, e.g.
"/home/djangoprojects/myproject".
--traceback Raise on CommandError exceptions
--no-color Don't colorize the command output.
--force-color Force colorization of the command output.
--skip-checks Skip system checks.
With this utility, it is possible to separately launch clustering of faces, bodies, and vehicles and perform force clustering when the maximum number of cluster events exceeds the value of the CLUSTERIZATION_MAX_CLUSTER_EVENTS parameter (see Enable and Configure Clustering). For example, to force start face clustering, execute:
sudo docker exec -it findface-multi-findface-multi-legacy-1 /opt/findface-security/bin/python3 /tigre_prototype/manage.py run_clusterization --object-types face --force
Reports
In this chapter:
Report Types
The following reports are available:
Standard reports on the system entities:
Work time reports, that allow you to regard the exact moments of staff entering and exiting an enterprise or a designated area and calculate the total time spent at it, can be built on the Events tab.
Build Standard Report
To build a standard report on a system entity, do the following:
Navigate to the tab associated with the required entity: Search, Episodes & Events, Clusters, Video Sources, Record index, Audience analysis, Audit Logs.
Click the More filters button. Set filters for the report.
Click Apply and create report.
Specify the report name.
If you are on the Events tab, select the report type Standard report (as there are two types of reports available).
If applicable, choose whether to save the report images as links, thumbnails, or full frames.
Click Add. The report will be available for download on the Reports tab.
Build Work Time Report
A work time report can be built on the Events tab. It requires specifying entrance and exit cameras.
To build a work time report, do the following:
Set filters for the report.
Click Apply and create report.
Specify the report name.
Select the report format.
Select the report type: Work time report.
If applicable, choose whether to save the report images as links, thumbnails, or full frames.
Select the entrance camera.
Select the exit camera.
Click Add. The report will be available for download on the Reports tab.
Work with Reports
You can access reports previously created in the system on the Reports tab. The following operations are available:
Download selected reports.
Update selected reports.
Delete selected reports.
See also
Audit Log
The FindFace Multi comprehensive and searchable audit log is an excellent complementary tool for user management that provides you with a thorough audit of the user actions and strengthens your system protection. You can access this functionality on the Audit Log tab.
Each record provides the following data:
username of the user who performed the action
IP address where the request came from
device id: the unique identifier of the client device
action type such as authorization, search, object modification, restart, and so on
object type to which the action applies, for example, a record
object identifier
details, subject to the action type
timestamp
Use the filter panel above to set up the search conditions.
People-Related Analytics
FindFace Multi is an ideal tool to gather people-related analytics. Enable face and body clusters first and then make the most of them with our analytical features.
Interaction Tracking
FindFace Multi can provide interaction tracking and analysis. To harness this functionality, you must enable face clustering.
When tracking interactions, the system automatically determines a circle of people with whom a person has previously been in contact. For each person from the first circle, the system builds another circle of connected people, and so on. Overall, interaction analysis is three-circles deep.
Note
You can track interactions between bodies and vehicles as well. Access this functionality via HTTP API. Note that body and vehicle clustering must be enabled beforehand.
The interaction analysis is available on the Interaction Tracking tab.
Tip
You can also display the circle of connected people right from the Clusters tab and Record index tab by clicking on the handshake icon.
On the Interaction Tracking tab, click on a person to display their first circle of interactions and associated events. Keep on to unveil the entire tree of interactions.
You can apply the available filters to the interaction tracking:
Tip
For example, you can find older adults or people without a face mask who are directly or indirectly related to a potentially contagious person.
Matches: display clusters only with/without matches, or all events.
Watch lists: display clusters only for a selected watch list.
Camera groups: display only clusters from a selected group of cameras.
Cameras: display only clusters from a selected camera.
Record name: display clusters with a given record name.
Date and time: display only clusters formed within a certain period.
First cluster even: display only the first cluster event within a certain period.
Cluster event: display only cluster event within a certain period.
ID: display a cluster with a given ID.
Specific filters for face clusters
Age: display clusters with people of a given age.
Beard: filter clusters by the fact of having a beard.
Emotions: display clusters with given emotions.
Gender: display clusters with people of a given gender.
Glasses: filter clusters by the fact of wearing glasses.
Liveness: filter clusters by face liveness.
Face mask: filter clusters by the fact of wearing a face mask.
When searching through a circle of interactions, apply the following additional settings:
Interaction search based on the episode last event: check to analyze contacts between individuals using the last event of an episode. In this case, having found truly associated people is most probable as they simultaneously leave a camera’s field of view. If the option is disabled, the system will use the best event of an episode for interaction search.
Maximum gap between the appearance of individuals to consider them connected, seconds: maximum time in seconds between the appearance of individuals to consider them related.
Audience analysis
FindFace Multi includes statistics on the number of visitors, their gender, average age, most frequently visited zones (judged by most active cameras), and the character of visits (first-timers or returners).
The analytical data charts are available on the Audience analysis tab.
Important
The analytics are built only when the face clustering is enabled.
To work with the analytical data, use the following filters:
Date and time
Scale interval
Minimum time between visits
Number of returns
Cameras
Watch lists
Age
Gender
See also
Video Surveillance
In this chapter:
Video Wall
The FindFace Multi Video Wall allows for basic video surveillance. Use it to display the video image from cameras and video files.
Important
Advanced video surveillance is available with Video Recorder.
Display Video
The Video Wall offers two modes, four predefined layouts in each:
video streaming
video streaming with object detection and episode feed
To display video on the Video Wall, do the following:
Navigate to the Video Wall tab.
Select a mode and camera layout.
Drag-n-drop cameras of your choice to the Video Wall.
You can work with the episode and event feed on the Video Wall in the same manner as with the Episodes & Events tab, including the same filters that are available for episodes and events.
Note
See the full list of episode filters and event filters applicable to the Video Wall.
Video Recorder
Enable Video Recording
If a Video Recorder is deployed and configured, you will see the Record video checkbox appear in the main camera settings. Check it to enable recording video chunks from a camera and sending them over to Video Recorder for further processing.
View Camera Video in Video Player
Under specific settings, clicking on the camera preview opens up not a static frame, but a video player, streaming live video from the camera.
The video player has a highly intuitive design. It provides the following possibilities:
Stream live video from the camera.
Watch video chunks recorded from the camera. They will be marked in purple on the timeline. To quickly switch from archived video to the camera live stream, click the Live button.
Visual indication of moments corresponding to face or vehicle events (if enabled) as colorful markers on the timeline. The unmatched events are marked in pink, while the matched ones are light green.
Note
This feature is optional and must be pre-configured. See Configure Video Recorder.
Important
The time on the timeline is displayed in the spectator’s time zone. For example, if an event happened at 2 p.m. in Abu Dhabi, its marker would be at 12 p.m. for a spectator in Paris.
Navigate to the past and future over the timeline, with the possibility of zooming it in and out.
Tip
You can navigate along the timeline by using the < / > buttons, or by moving the cursor along the timeline with the right mouse button held down.
Tip
To zoom in and out, use the mouse wheel with Ctrl.
To export selected clips, select the interval and click Download.
View Video from the Event and Episode
You may view video from the episode or event of your interest, if you have enabled Video Recorder and if you have a recorded video archive during the event or episode. When hovering, the play button will display on the thumbnail of the object (face/body or vehicle). Press it to start playing the video from the time when the event/episode was recorded from the camera.
Integrations
This chapter is all about integration with FindFace. In the current version of FindFace, you can only integrate your system via HTTP API.
HTTP API
Detailed interactive documentation on the FindFace HTTP API is available after installation at http://<findface-ip:port>/api-docs. Learn and try it out.
Tip
<findface-ip:port> must be substituted into your ip address for FindFace Multi.
Tip
You can also find it by navigating to in the web interface.
Overview
Tip
If after having read this section, you still have questions, do not hesitate to contact our experts by email: support@ntechlab.com.
Authentication
All API methods require a simple token-based HTTP Authentication. To authenticate, put the word “Token” and a token, separated by whitespace, into the Authorization HTTP header: Authorization: Token 000... All requests that fail to provide a valid authentication token will result in an HTTP 401 Unauthorized response.
Parameters Format
There are two ways to pass parameters to the API methods:
application/json: parameters are represented by a JSON contained in the body.multipart/form-data: parameters are encoded into separate parts. This way supports uploading a photo image file in the same request.
Additional Information
Standard extraction limits:
Image formats: JPEG, PNG, WEBP
Maximum photo file size: 10 MB
Maximum photo resolution: 6000 pixels on the biggest side
Minimal size of a face: 50x50 pixels
Check /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml for custom definition.
Getting started
In this chapter, you can find a recommended sequence of steps that will help you harness your system’s complete functionality.
In this chapter:
Tip
<findface-ip:port> must be substituted into your ip address for FindFace Multi.
<token> must be substituted into your token.
Gear Up for Work
Perform the primary configuration of your system:
Deploy FindFace Multi and learn API documentation at
http://<findface-ip:port>/api-docs.Authorize with HTTP Basic auth by sending credentials to the login form.
Create an authorization token with your session UUID and optionally face authorization.
Paste a token into the
API Keyfield in the authorization form:Token <token>or put it into the Authorization HTTP header.
Organize Watch Lists and Card Index
Create a new watch list, using
POSTmethod, or use the default one.Find watch list ID to make a
POSTrequest for creating a new card.Perform face (car, body) detection on the attached photo and copy
idof the necessary object.Attach an image of the individual’s face, using
detection id.
Upload and Process Video File
Organize Video Surveillance by uploading a video file.
Find
idof the video archive by listing video archives or adding new one.Attach source video file with
idof the video archive.
Data Sources
To configure video-based object monitoring, add cameras to FindFace Multi, grouping them subject to their location.
A default preconfigured camera group is available in the system. List available camera group.
If necessary, create a new camera group.
Use camera group
idto make aPOSTrequest for creating a new camera.
FindFace Multi in Action
Search for objects in the database of detected objects and card index. Learn more.
Compare two objects and verify that they match.
Authentication
API requests to FindFace Multi are to be sent to http://<findface-ip:port>.
HTTP Basic authentication
HTTP basic authentication is a simple challenge and response mechanism with which a server can request authentication information (a username and password).
Put your credentials to the login form:
Username – username of your FindFace Multi account.
Password – password for your FindFace Multi account.
This data are to be sent in the Authorization request header as base64 code.
Authorization: Basic <base64(username:password)>
API Key authorization
To create an authorization token, use the following method:
POST /auth/login/
You should provide basic authorization credentials (username, password) earlier.
This method returns a structure containing an authorization token, that you can use in all other methods.
If the system uses face and password authorization, together with the Basic authorization header, additionally pass the token received during face authorization in the video_auth_token field.
The REQUEST BODY is required and contains application/json object with the following parameters:
Tip
* - means required parameters. 🆁 – read only.
Option |
Schema |
Description |
|---|---|---|
|
string |
Constraints: Min 1 chars. Inactive token from face authorization. Required when |
|
string |
Constraints: 1 to 256 chars. Session unique identifier on the device. |
|
boolean |
Device is mobile. |
|
<any-key>: str┃int┃float┃bool┃object┃array┃null |
Device information. |
Request example
{
"video_auth_token": "A",
"uuid": "A",
"mobile": false,
"device_info": {}
}
CURL Example
curl -X POST "http://<findface-ip:port>/auth/login/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Basic <base64(username:password)>" \
-H "Content-Type: application/json" \
-d '{"uuid":"A"}' \
Returns:
200 on success.
401 if response body is unspecified.
If the response is successful (OK: 200), it returns an object that contains the following parameters. Example.
Response example
Option |
Schema |
Description |
|---|---|---|
|
string |
Authorization token. |
|
{…} |
User info. |
|
date-time |
Token expiration datetime. |
user contains the following parameters:
id*: integer 🆁
active*: boolean
created_date*: date-time 🆁
groups: [integer]
modified_date*: date-time 🆁 – Object modification date.
permissions*: [string]
real_name*: string – Constraints: Max 80 chars.
name*: string – Constraints: Max 80 chars.
comment: string – Extended description of the user, up to 2048 chars.
camera_group_permissions*: {<any-key>: string}
watch_list_permissions*: {<any-key>: string}
group_permissions*: {<any-key>: string}
primary_group*: integer
language: enum – Allowed: en-us┃es┃ru. Backend message language for the user.
has_face*: boolean 🆁 – User photo is provided.
face_cover: string┃null – Constraints: Max 32 chars.
ad_user*: boolean 🆁 – User is registered in Active Directory.
Response example
{
"token": "string",
"user": {
"id": 0,
"active": false,
"created_date": "1970-01-01T00:00:00.000Z",
"groups": [
0
],
"modified_date": "1970-01-01T00:00:00.000Z",
"permissions": [
"string"
],
"real_name": "AAAAAA",
"name": "AAAAAA",
"comment": "AAAAAA",
"camera_group_permissions": {},
"watch_list_permissions": {},
"group_permissions": {},
"primary_group": 0,
"language": "en-us",
"has_face": false,
"face_cover": "AAAAAA",
"ad_user": false
},
"token_expiration_datetime": "1970-01-01T00:00:00.000Z"
}
Paste this obtained token into the API Key field in the authorization form: Token <token> and click SET or put it into the Authorization HTTP header: Authorization: Token 000.... All requests that fail to provide a valid authentication token will result in an HTTP 401 Unauthorized response.
To disable applied keys, click REMOVE or CLEAR ALL API KEYS or send a request POST /auth/logout/.
Watch lists
List watch lists
To list watch lists, use the following method:
GET /watch-lists/
The REQUEST contains the following QUERY-STRING PARAMETERS:
Name |
Schema |
Description |
|---|---|---|
|
integer |
Number of results to return. |
|
string |
Available fields: id, created_date, modified_date. |
To look for watch lists id point out some value of limit and for ordering.
CURL example
curl -X GET "http://<findface-ip:port>/watch-lists/?limit=2&ordering=id" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
If the response is successful (OK: 200), it returns an object that contains array of results with the following parameters. Example.
Tip
* – means required parameters. 🆁 – read only.
Name |
Schema |
Description |
|---|---|---|
|
integer 🆁 |
Watch list ID. |
|
date-time 🆁 |
Object creation date. |
|
date-time 🆁 |
Object modification date. |
|
boolean |
|
|
string |
Short watch list name, up to 256 characters. |
|
string |
Extended description, up to 2048 characters. |
|
string |
Color of the object label in hex. Constraints: Max 6 chars. |
|
boolean |
|
|
boolean |
Require manual acknowledgment of the events that match with the watch list. |
|
{<any-key>: string} |
Permissions. |
|
[integer] |
List of the camera groups used to monitor the watch list. |
|
number┃null |
Watch list face confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Watch list body confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Watch list car confidence threshold. Constraints: Min 0┃Max 1. |
|
boolean |
Events won’t be created if set to |
|
boolean |
|
|
date-time┃null |
Data-time information. |
|
date-time┃null |
Data-time information. |
|
{…} |
Serializer mixin that raises ValidationError if excess fields are presented. Can be used in nested serializers. |
|
date-time |
Data-time information. |
|
string |
Constraints: Max 256 chars bit. |
Response example
{
"results": [
{
"id": -1,
"created_date": "2023-01-16T13:44:36.407610Z",
"modified_date": "2023-01-16T13:44:36.407750Z",
"active": true,
"name": "Unmatched",
"comment": "Default list for unmatched events",
"color": "ffffff",
"notify": false,
"acknowledge": false,
"permissions": {
"1": "edit",
"2": "view",
"3": "view"
},
"camera_groups": [],
"face_threshold": null,
"body_threshold": null,
"car_threshold": null,
"ignore_events": false,
"send_events_to_external_vms": false,
"active_after": null,
"active_before": null,
"disable_schedule": {},
"recount_schedule_on": null,
"origin": "ffsecurity"
},
{
"id": 1,
"created_date": "2023-01-16T13:44:36.394542Z",
"modified_date": "2023-01-16T13:44:36.394577Z",
"active": true,
"name": "Default Watch List",
"comment": "",
"color": "123456",
"notify": false,
"acknowledge": false,
"permissions": {
"1": "edit",
"2": "view",
"3": "view"
},
"camera_groups": [],
"face_threshold": null,
"body_threshold": null,
"car_threshold": null,
"ignore_events": false,
"send_events_to_external_vms": false,
"active_after": null,
"active_before": null,
"disable_schedule": {},
"recount_schedule_on": null,
"origin": "ffsecurity"
}
]
}
Create a new watch list
To add a new watch list, use the following method:
POST /watch-lists/
The REQUEST BODY is required and contains application/json object with the watch lists parameters.
Name |
Schema |
Description |
|---|---|---|
|
boolean |
|
|
string |
Short watch list name, up to 256 characters. |
|
string |
Extended description, up to 2048 characters. |
|
string |
Color of the object label in hex. Constraints: 1 to 6 chars. |
|
boolean |
|
|
boolean |
Require manual acknowledgment of the events that match with the watch list. |
|
<any-key>: string |
Constraints: Min 1 chars. |
|
[integer] |
List of the camera groups used to monitor the watch list. |
|
number┃null |
Watch list face confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Watch list body confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Watch list car confidence threshold. Constraints: Min 0┃Max 1. |
|
boolean |
Events won’t be created if set to |
|
boolean |
|
|
date-time┃null |
Date-time information. |
|
date-time┃null |
Date-time information. |
|
{…} |
Serializer mixin that raises ValidationError if excess fields are presented. Can be used in nested serializers. |
|
string |
Constraints: 1 to 256 chars bit. |
Request example
Tip
This example is given for reference only, substitute your values in the corresponding fields. You may fill in only the required fields, and the others will be by default.
{
"active": false,
"name": "A",
"comment": "AAAAAA",
"color": "A",
"notify": false,
"acknowledge": false,
"permissions": {},
"camera_groups": [
0
],
"face_threshold": 0,
"body_threshold": 0,
"car_threshold": 0,
"ignore_events": false,
"send_events_to_external_vms": false,
"active_after": "1970-01-01T00:00:00.000Z",
"active_before": "1970-01-01T00:00:00.000Z",
"disable_schedule": {
"monday": [
[
"A"
]
],
"tuesday": [
[
"A"
]
],
"wednesday": [
[
"A"
]
],
"thursday": [
[
"A"
]
],
"friday": [
[
"A"
]
],
"saturday": [
[
"A"
]
],
"sunday": [
[
"A"
]
]
},
"origin": "A"
}
You may send these parameters:
{
"active": true,
"name": "Test_list",
"comment": "AAAAAA",
"color": "35a2ee"
}
CURL example
curl -X POST "http://<findface-ip:port>/watch-lists/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: application/json" \
-d '{"active":true,"name":"Test_list","comment":"AAAAAA","color":"35a2ee"}' \
If the response is successful (Created: 201), it returns an object that contains parameters. Example.
Response example
{
"id": 2,
"created_date": "2023-01-16T14:07:52.424520Z",
"modified_date": "2023-01-16T14:07:52.424549Z",
"active": true,
"name": "Test_list",
"comment": "AAAAAA",
"color": "35a2ee",
"notify": false,
"acknowledge": false,
"permissions": {
"1": "edit"
},
"camera_groups": [],
"face_threshold": 0,
"body_threshold": 0,
"car_threshold": 0,
"ignore_events": false,
"send_events_to_external_vms": false,
"active_after": null,
"active_before": null,
"disable_schedule": {},
"recount_schedule_on": null,
"origin": "ffsecurity"
}
Use watch list id to make a POST request for creating a new card.
Useful requests
GET /watch-lists/
POST /watch-lists/
GET /watch-lists/{id}/
DELETE /watch-lists/{id}/
PATCH /watch-lists/{id}/
POST /watch-lists/{id}/purge/
GET /watch-lists/count/
POST /watch-lists/purge_all/
Record
Create a new record
To create a new human record, use the following method:
POST /cards/humans/
Tip
For vehicle records instead humans use cars.
The REQUEST BODY is required and contains application/json object with the following parameters.
Tip
* – means required parameters. 🆁 – read only.
Name |
Schema |
Description |
|---|---|---|
|
boolean |
|
|
string |
Record name, up to 256 characters. |
|
string |
Extended description, up to 2048 characters. |
|
[integer] |
Array of the related Watch lists ID. |
|
<any-key> |
Meta data. |
|
date-time┃null |
Date-time information. |
|
date-time┃null |
Date-time information. |
|
{…} |
Serializer mixin that raises ValidationError if excess fields are presented. Can be used in nested serializers. |
Request example
Tip
This example is given for reference only, substitute your values in the corresponding fields. You may fill in only the required fields, and the others will be by default.
{
"active": false,
"name": "A",
"comment": "AAAAAA",
"watch_lists": [
0
],
"meta": {},
"active_after": "1970-01-01T00:00:00.000Z",
"active_before": "1970-01-01T00:00:00.000Z",
"disable_schedule": {
"monday": [
[
"A"
]
],
"tuesday": [
[
"A"
]
],
"wednesday": [
[
"A"
]
],
"thursday": [
[
"A"
]
],
"friday": [
[
"A"
]
],
"saturday": [
[
"A"
]
],
"sunday": [
[
"A"
]
]
}
}
For example, you may send these parameters:
{
"active": true,
"name": "Angelina Jolie",
"comment": "Миссис Смит",
"watch_lists": [
2
]
}
CURL example
curl -X POST "http://<findface-ip:port>/cards/humans/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: application/json" \
-d '{"active":true,"name":"Angelina Jolie","comment":"Миссис Смит","watch_lists":[2]}' \
If the response is successful (Created:201), it returns an object that contains the following parameters. Example.
Name |
Schema |
Description |
|---|---|---|
|
integer 🆁 |
Record ID. |
|
boolean |
|
|
boolean 🆁 |
|
|
date-time 🆁 |
Object creation date. |
|
date-time 🆁 |
Object modification date. |
|
string |
Record name, up to 256 characters. |
|
string |
Extended description, up to 2048 characters. |
|
[integer] |
Array of the related Watch lists ID. |
|
|
|
|
<any-key> |
Meta data. |
|
number 🆁 |
Looks like confidence. |
|
date-time┃null |
Data-time information. |
|
date-time┃null |
Data-time information. |
|
array |
Serializer mixin that raises ValidationError if excess fields are presented. Can be used in nested serializers. |
|
date-time 🆁 |
Data-time information. |
|
integer 🆁 |
Linked face objects counter. |
|
integer 🆁 |
Linked body objects counter. |
|
integer 🆁 |
Face claster. |
|
integer 🆁 |
Body cluster. |
|
|
|
Note
Relations are not supported in FindFace Multi 2.1 and are kept in API only.
Response example
{
"id": 1,
"active": true,
"filled": true,
"created_date": "2023-01-17T07:43:56.363330Z",
"modified_date": "2023-01-17T07:43:56.363354Z",
"name": "Angelina Jolie",
"comment": "Миссис Смит",
"watch_lists": [
2
],
"meta": {},
"active_after": null,
"active_before": null,
"disable_schedule": {},
"recount_schedule_on": null,
"face_objects": 0,
"body_objects": 0,
"face_cluster": null,
"body_cluster": null,
"links_to_relations": []
}
List records
To list human records, use this method:
GET /cards/humans/
The REQUEST contains the QUERY-STRING PARAMETERS. All parameters you can see in http://<findface-ip:port>/api-docs.
You may use this method to search faces in the system. Specify the value of detection:<detection id>, received by detection of objects on a photo, in the field looks_like.
CURL example
curl -X GET "http://<findface-ip:port>/cards/humans/?looks_like=detection%3Acf4ffvmv54rotim9jt60" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
If the response is successful (OK: 200), it returns an object that contains the following parameters:
Name |
Schema |
Description |
|---|---|---|
|
string┃null |
Next page. |
|
string┃null |
Previous page. |
|
[{…}] |
Contain the following parameters. |
Response example
{
"next_page": null,
"prev_page": null,
"results": [
{
"id": 1,
"active": true,
"filled": true,
"created_date": "2023-01-17T07:43:56.363330Z",
"modified_date": "2023-01-17T11:56:42.496871Z",
"name": "Angelina Jolie",
"comment": "Миссис Смит",
"watch_lists": [
2
],
"looks_like": {
"confidence": 0.8078,
"collection": "face_objects",
"matched_object": "4493493039043981648"
},
"meta": {},
"looks_like_confidence": 0.8078,
"active_after": null,
"active_before": null,
"disable_schedule": {},
"recount_schedule_on": null,
"face_objects": 1,
"body_objects": 0,
"face_cluster": 11,
"body_cluster": null,
"links_to_relations": []
}
]
}
Useful requests
GET /cards/cars/
POST /cards/cars/
GET /cards/cars/{id}/
DELETE /cards/cars/{id}/
PATCH /cards/cars/{id}/
GET /cards/cars/count/
GET /cards/humans/
POST /cards/humans/
GET /cards/humans/{id}/
DELETE /cards/humans/{id}/
PATCH /cards/humans/{id}/
GET /cards/humans/count/
Detect objects on a photo
To detect objects on a photo, use this method:
POST /detect
The REQUEST BODY is required and contains multipart/form-data with the following parameters.
Name |
Schema |
Description |
|---|---|---|
|
binary |
Source image file. |
|
object |
Attributes of |
attributes might be empty or contain the objects of face, car and body with the following parameters for each one:
age: boolean
beard: boolean
emotions: boolean
glasses: boolean
gender: boolean
medmask: boolean
headpose: boolean
Attach a source image file and send the POST request.
CURL example
curl -X POST "http://<findface-ip:port>/detect" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: multipart/form-data" \
-F "photo=@Смит_01.png" \
-F "attributes={
"face": {
"age": false,
"beard": false,
"emotions": false,
"glasses": false,
"gender": false,
"medmask": false,
"headpose": false
},
"car": {
"description": false,
"license_plate": false,
"special_vehicle_type": false,
"category": false,
"weight_type": false,
"orientation": false
},
"body": {
"color": false,
"clothes": false,
"bags": false,
"protective_equipment": false,
"age_gender": false
}
}" \
If the response is successful (OK: 200), it returns an object that contains the following parameters. Example.
Name |
Schema |
Description |
|---|---|---|
|
integer |
EXIF orientation of the photo. |
|
<any-key>: str┃int┃float┃bool┃object┃array┃null |
Returned objects with requested attributes. |
Response example
{
"orientation": 1,
"objects": {
"face": [
{
"id": "cf0mbqev54rqhngnq940",
"bbox": {
"left": 451,
"top": 235,
"right": 645,
"bottom": 502
},
"detection_score": 0.80645436,
"low_quality": false,
"features": {}
},
{
"id": "cf0mbqev54rqhngnq94g",
"bbox": {
"left": 757,
"top": 79,
"right": 948,
"bottom": 353
},
"detection_score": 0.90099674,
"low_quality": false,
"features": {}
}
]
}
}
In the response you will get ID of the object and the coordinates of bbox of the object (face, car, body). Copy the value of id of the necessary object to use it for adding an object to the card or to search an object.
Objects
Add face object
POST /objects/faces/
This method creates new Face object that contains source photo, thumbnail, and other attributes.
Tip
For car or body objects instead faces use cars or body.
The REQUEST BODY is required and contains multipart/form-data with the following parameters:
Name |
Schema |
Description |
|---|---|---|
|
string |
This field can contain one of the following references:
|
|
string |
Auxiliary parameter. |
|
enum |
This parameter defines the FindFace Multi default behaviour when there are multiple objects are present in
|
|
integer |
Add object to this upload list. |
|
binary |
Source photo (required when |
|
integer |
Left border of object’s bounding box. |
|
integer |
Top border of object’s bounding box. |
|
integer |
Right border of object’s bounding box. |
|
integer |
Bottom border of object’s bounding box. |
|
boolean |
|
|
integer |
Related |
Attach the photofile to the source_photo, point out the related card id and paste detection:<detection id> into create_from, where detection id you have got from detection objects on a photo. source_photo must contain the same image that was used for POST /detect.
CURL example
curl -X POST "http://<findface-ip:port>/objects/faces/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: multipart/form-data" \
-F "create_from=detection:cf2g86uv54rqhngnq960" \
-F "source_photo=@Смит_01.png" \
-F "card=3" \
If the response is successful (Created: 201), it returns an object that contains the following parameters. Example.
Name |
Schema |
Description |
|---|---|---|
|
string 🆁 |
ID. |
|
date-time 🆁 |
Object creation date. |
|
date-time 🆁 |
Object modification date. |
|
string 🆁 |
Filename supplied for source_photo on object creation. |
|
integer |
Left border of object’s bounding box. |
|
integer |
Top border of object’s bounding box. |
|
integer |
Right border of object’s bounding box. |
|
integer |
Bottom border of object’s bounding box. |
|
uri 🆁 |
Object thumbnail. |
|
boolean |
Default: true. |
|
<any-key>: str┃int┃float┃bool┃object┃array┃null |
Features. |
|
integer |
Related card ID. |
Response example
{
"card": 3,
"created_date": "2023-01-16T08:12:55+00:00",
"modified_date": "1970-01-01T00:00:00+00:00",
"source_photo_name": "Смит_01.png",
"source_photo": "http://172.23.218.94/uploads/cards/7w/3/face_%D0%A1%D0%BC%D0%B8%D1%82_01_ftgY5K.png",
"thumbnail": "http://172.23.218.94/uploads/cards/TD/3/face_%D0%A1%D0%BC%D0%B8%D1%82_01_thumbnail_VyVsIj.png",
"frame_coords_left": 757,
"frame_coords_top": 79,
"frame_coords_right": 948,
"frame_coords_bottom": 353,
"active": true,
"features": {},
"id": "4493225067924944019",
"meta": {}
}
Useful requests
GET /objects/bodies/
POST /objects/bodies/
GET /objects/bodies/{id}/
DELETE /objects/bodies/{id}/
PATCH /objects/bodies/{id}/
GET /objects/cars/
POST /objects/cars/
GET /objects/cars/{id}/
DELETE /objects/cars/{id}/
PATCH /objects/cars/{id}/
GET /objects/faces/
POST /objects/faces/
GET /objects/faces/{id}/
DELETE /objects/faces/{id}/
PATCH /objects/faces/{id}/
Videos
List video archives
To list video archives, use the following method:
GET /videos/
The REQUEST contains QUERY-STRING PARAMETERS.
Name |
Schema |
Description |
|---|---|---|
|
array of integer |
Select videos included in these cases. |
|
date-time |
Select objects with |
|
date-time |
Select objects with |
|
number |
Select objects with |
|
date-time |
Select objects with |
|
date-time |
Select objects with |
|
number |
Select objects with |
|
number |
Select objects with |
|
string |
Number of results to return per page. |
|
string |
Select video archives with this |
|
string |
Available fields: id, created_date, name. |
|
string |
Pagination cursor value. |
|
string |
Select video archives with this save_to field. |
For example, select some number of results for limit.
CURL example
curl -X GET "http://<findface-ip:port>/videos/?limit=2" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
If the response is successful (OK: 200), it returns an object that contains with the following parameters. Example.
Name |
Schema |
|---|---|
|
string┃null |
|
string┃null |
|
array |
The results contain an array with the following parameters:
Name |
Schema |
Description |
|---|---|---|
|
integer 🆁 |
ID of the video archives. |
|
integer |
Camera group. |
|
string┃null |
Name of the video archives. Constraints: Max 256 chars. |
|
string┃null |
URL. |
|
integer┃null |
Camera. |
|
date-time 🆁 |
Video processing start date. |
|
boolean 🆁 |
Processing is active. |
|
uri 🆁 |
Screenshot URL. |
|
{…} |
Serializer mixin that raises ValidationError if excess fields are presented. Can be used in nested serializers. |
|
number 🆁 |
Source length in seconds. |
|
{…} |
Additional status information. |
|
boolean 🆁 |
|
|
boolean 🆁 |
|
|
integer 🆁 |
Number of created faces. |
|
integer 🆁 |
Video file size in the archive. |
|
date-time 🆁 |
Object creation date. |
|
integer 🆁 |
Number of created bodies. |
|
integer 🆁 |
Number of created cars. |
|
integer┃null |
Case. |
|
integer┃null 🆁 |
Count of face clusters created from this video. |
|
integer┃null 🆁 |
Count of body clusters created from this video. |
|
integer┃null 🆁 |
Count of car clusters created from this video. |
Note
Cases are not supported in FindFace Multi 2.1 and are kept in API only.
Response example
{
"next_page": null,
"prev_page": null,
"results": [
{
"id": 2,
"camera_group": 1,
"name": "file.mp4",
"url": null,
"camera": null,
"processing_start_date": null,
"active": false,
"screenshot": "http://<findface-ip:port>/videos/2/screenshot/",
"stream_settings": {
"detectors": {
"face": {
"filter_max_size": 8192,
"filter_min_quality": 0.45,
"filter_min_size": 60,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 95,
"overall_only": true,
"post_best_track_frame": true,
"post_best_track_normalize": true,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 1,
"roi": "",
"track_interpolate_bboxes": true,
"track_max_duration_frames": 0,
"track_miss_interval": 1,
"track_overlap_threshold": 0.25,
"track_send_history": false,
"tracker_type": "simple_iou",
"track_deep_sort_matching_threshold": 0.65,
"track_deep_sort_filter_unconfirmed_tracks": true
},
"body": null,
"car": null
},
"disable_drops": true,
"ffmpeg_format": "",
"ffmpeg_params": [],
"imotion_threshold": 0,
"play_speed": -1,
"rot": "",
"router_timeout_ms": 15000,
"router_verify_ssl": true,
"start_stream_timestamp": 0,
"stream_data_filter": "",
"use_stream_timestamp": false,
"video_transform": "",
"enable_recorder": false,
"enable_liveness": false
},
"source_len": null,
"health_status": {
"enabled": false,
"status": "DISABLED",
"msg": "",
"statistic": {
"processed_duration": 0,
"faces_posted": 0,
"faces_failed": 0,
"faces_not_posted": 0,
"processing_fps": 0,
"frames_dropped": 0,
"frames_processed": 0,
"frames_imotion_skipped": 0,
"decoding_soft_errors": 0,
"frame_width": 0,
"frame_height": 0,
"last_stream_timestamp": 0,
"objects": null,
"job_starts": 0
},
"code": "gray",
"code_desc": "Обработка видео не запущена"
},
"finished": false,
"queued": false,
"face_count": 0,
"file_size": 2259950,
"created_date": "2023-01-12T08:57:36.811305Z",
"body_count": 0,
"car_count": 0,
"case": null,
"face_cluster_count": 0,
"body_cluster_count": 0,
"car_cluster_count": 0
}
]
}
Add a new video archive
To add video archives, use the following method:
POST /videos/
The REQUEST BODY is required and contains application/json object with the following parameters:
Name |
Schema |
Description |
|---|---|---|
|
integer |
ID value of camera group. |
|
string┃null |
Name of the video archive. Constraints: 1 to 256 chars. |
|
string┃null |
Constraints: Min 1 chars. |
|
integer┃null |
Camera. |
|
{…} |
Serializer mixin that raises ValidationError if excess fields are presented. Can be used in nested serializers. |
|
integer┃null |
Case. |
Note
Cases are not supported in FindFace Multi 2.1 and are kept in API only.
Request example
Tip
This example is given for reference only, substitute your values in the corresponding fields. You may fill in only the required fields, and the others will be by default.
{
"camera_group": 0,
"name": "A",
"url": "A",
"camera": 0,
"stream_settings": {
"detectors": {
"face": {
"filter_max_size": 0,
"filter_min_quality": 0,
"filter_min_size": 0,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 0,
"overall_only": false,
"post_best_track_frame": false,
"post_best_track_normalize": false,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 0,
"roi": "string",
"track_interpolate_bboxes": false,
"track_max_duration_frames": 0,
"track_miss_interval": 0,
"track_overlap_threshold": 0,
"track_send_history": false,
"tracker_type": "string",
"track_deep_sort_matching_threshold": 0,
"track_deep_sort_filter_unconfirmed_tracks": false
},
"body": {
"filter_max_size": 0,
"filter_min_quality": 0,
"filter_min_size": 0,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 0,
"overall_only": false,
"post_best_track_frame": false,
"post_best_track_normalize": false,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 0,
"roi": "string",
"track_interpolate_bboxes": false,
"track_max_duration_frames": 0,
"track_miss_interval": 0,
"track_overlap_threshold": 0,
"track_send_history": false,
"tracker_type": "string",
"track_deep_sort_matching_threshold": 0,
"track_deep_sort_filter_unconfirmed_tracks": false
},
"car": {
"filter_max_size": 0,
"filter_min_quality": 0,
"filter_min_size": 0,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 0,
"overall_only": false,
"post_best_track_frame": false,
"post_best_track_normalize": false,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 0,
"roi": "string",
"track_interpolate_bboxes": false,
"track_max_duration_frames": 0,
"track_miss_interval": 0,
"track_overlap_threshold": 0,
"track_send_history": false,
"tracker_type": "string",
"track_deep_sort_matching_threshold": 0,
"track_deep_sort_filter_unconfirmed_tracks": false
}
},
"disable_drops": false,
"ffmpeg_format": "string",
"ffmpeg_params": [
"A"
],
"imotion_threshold": 0,
"play_speed": 0,
"rot": "string",
"router_timeout_ms": 0,
"router_verify_ssl": false,
"start_stream_timestamp": 0,
"stream_data_filter": "string",
"use_stream_timestamp": false,
"video_transform": "string",
"enable_recorder": false,
"enable_liveness": false
},
"case": 0
}
For example, you may send these parameters.
{
"camera_group": 1,
"name": "Pitt&Jolie"
}
CURL example
curl -X POST "http://<findface-ip:port>/videos/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: application/json" \
-d '{"camera_group":1,"name":"Pitt&Jolie"}' \
If the response is successful (Created: 201), it returns an object that contains parameters. Example.
Response example
{
"id": 6,
"camera_group": 1,
"name": "Pitt&Jolie",
"url": null,
"camera": null,
"processing_start_date": null,
"active": false,
"screenshot": "http://<findface-ip:port>/videos/6/screenshot/",
"stream_settings": {
"detectors": {
"face": {
"filter_max_size": 8192,
"filter_min_quality": 0.45,
"filter_min_size": 60,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 95,
"overall_only": true,
"post_best_track_frame": true,
"post_best_track_normalize": true,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 1,
"roi": "",
"track_interpolate_bboxes": true,
"track_max_duration_frames": 0,
"track_miss_interval": 1,
"track_overlap_threshold": 0.25,
"track_send_history": false,
"tracker_type": "simple_iou",
"track_deep_sort_matching_threshold": 0.65,
"track_deep_sort_filter_unconfirmed_tracks": true
},
"body": {
"filter_max_size": 8192,
"filter_min_quality": 0.6,
"filter_min_size": 70,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 95,
"overall_only": true,
"post_best_track_frame": true,
"post_best_track_normalize": true,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 1,
"roi": "",
"track_interpolate_bboxes": true,
"track_max_duration_frames": 0,
"track_miss_interval": 1,
"track_overlap_threshold": 0.25,
"track_send_history": false,
"tracker_type": "simple_iou",
"track_deep_sort_matching_threshold": 0.65,
"track_deep_sort_filter_unconfirmed_tracks": true
},
"car": {
"filter_max_size": 8192,
"filter_min_quality": 0.73,
"filter_min_size": 100,
"fullframe_crop_rot": false,
"fullframe_use_png": false,
"jpeg_quality": 95,
"overall_only": true,
"post_best_track_frame": true,
"post_best_track_normalize": true,
"post_first_track_frame": false,
"post_last_track_frame": false,
"realtime_post_every_interval": false,
"realtime_post_first_immediately": false,
"realtime_post_interval": 1,
"roi": "",
"track_interpolate_bboxes": true,
"track_max_duration_frames": 0,
"track_miss_interval": 1,
"track_overlap_threshold": 0.25,
"track_send_history": false,
"tracker_type": "simple_iou",
"track_deep_sort_matching_threshold": 0.65,
"track_deep_sort_filter_unconfirmed_tracks": true
}
}
},
"source_len": null,
"health_status": {
"enabled": false,
"status": "WAITING_FOR_SYNC",
"msg": "",
"statistic": {},
"code": "red",
"code_desc": "Отсутствует задача в видео менеджере. Дождитесь синхронизации."
},
"finished": false,
"queued": false,
"face_count": 0,
"file_size": 0,
"created_date": "2023-01-18T08:38:52.119129Z",
"body_count": 0,
"car_count": 0,
"case": null,
"face_cluster_count": 0,
"body_cluster_count": 0,
"car_cluster_count": 0
}
Then point its id to upload the source file with PUT /videos/{id}/upload/source_file/.
Upload video file
To upload a video file, use the following method:
PUT /videos/{id}/upload/source_file/
The request contains required PATH PARAMETERS:
Name |
Schema |
Description |
|---|---|---|
|
integer |
A unique integer value identifying this video. |
In REQUEST BODY attach your video source file with id of the video archive.
Returns:
Created: 201 – in success.
Not Found: 404 – if fail.
CURL example
curl -X PUT "http://<findface-ip:port>/videos/6/upload/source_file/" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: video/mp4" \
--data-binary @pittjolie.mp4 \
Start video archive processing
To start video archive processing, use the following method:
POST /videos/{id}/process/
The REQUEST contains id in PATH PARAMETERS, identifying unique integer value of the video archive.
CURL example
curl -X POST "http://<findface-ip:port>/videos/6/process/" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
Returns:
OK: 200 – in success.
Not Found: 404 – if fail.
Useful requests
GET /videos/
POST /videos/
GET /videos/{id}/
PUT /videos/{id}/
DELETE /videos/{id}/
POST /videos/{id}/process/
GET /videos/{id}/screenshot/
POST /videos/{id}/screenshot/
POST /videos/{id}/stop/
PUT /videos/{id}/upload/source_file/
GET /videos/count/
Camera groups
Camera Group is a collection of Cameras, set up in same location or used for similar purposes.
List camera groups
To list camera groups, use the following method:
GET /camera-groups/
The REQUEST contains the following QUERY-STRING PARAMETERS:
Name |
Schema |
Description |
|---|---|---|
|
integer |
Number of results to return. |
|
string |
Available fields: |
To look for camera groups id specify some value of limit and ordering.
CURL example
curl -X GET "http://<findface-ip:port>/camera-groups/?limit=3&ordering=id" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
If the response is successful (OK: 200), it returns an object that contains array of results with the following parameters. Example.
Tip
* – means required parameters. 🆁 – read only.
Name |
Schema |
Description |
|---|---|---|
|
integer 🆁 |
ID of the camera group. |
|
date-time 🆁 |
Object creation date. |
|
date-time 🆁 |
Object modification date. |
|
boolean |
Process events from this camera group. |
|
string |
Short group name, up to 256 characters. |
|
string |
Extended description of the user, up to 2048 chars. |
|
boolean |
|
|
integer |
Event deduplication timeout. Constraints: Min 0 ┃ Max 10000. Default value is 15. |
|
<any-key>: string |
Labels used to allocate a certain |
|
<any-key>: string |
Permissions on this camera group. |
|
number┃null |
Camera group face confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Camera group body confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Camera group car confidence threshold. Constraints: Min 0┃Max 1. |
Response example
{
"results": [
{
"id": -1,
"created_date": "2023-01-16T13:44:36.531501Z",
"modified_date": "2023-01-16T13:44:36.531519Z",
"active": true,
"name": "Video archive default Camera Group",
"comment": "",
"deduplicate": false,
"deduplicateDelay": 15,
"labels": {},
"permissions": {
"1": "edit",
"2": "view",
"3": "view"
},
"face_threshold": null,
"body_threshold": null,
"car_threshold": null
}
]
}
Create Camera Group
To add a new camera group, use the following method:
POST /camera-groups/
The REQUEST BODY is required and contains application/json object with the following parameters. Example.
Name |
Schema |
Description |
|---|---|---|
|
boolean |
Process events from this camera group. |
|
string |
Short group name, up to 256 characters. |
|
string |
Extended description of the user, up to 2048 chars. |
|
boolean |
|
|
integer |
Event deduplication timeout. Constraints: Min 0┃Max 10000. Default value is 15. |
|
<any-key>: string |
Labels used to allocate a certain findface-video-worker instance to process video streams from this camera group. |
|
<any-key>: string |
Permissions on this camera group. |
|
number┃null |
Camera group face confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Camera group body confidence threshold. Constraints: Min 0┃Max 1. |
|
number┃null |
Camera group car confidence threshold. Constraints: Min 0┃Max 1. |
CURL example
curl -X POST "http://<findface-ip:port>/camera-groups/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: application/json" \
-d '{"active":false,"name":"Test_camera_group","comment":"AAAAAA","deduplicate":false,"deduplicateDelay":0,"labels":{},"permissions":{},"face_threshold":0,"body_threshold":0,"car_threshold":0}' \
Request example
Tip
This example is given for reference only, substitute your values in the corresponding fields. You may fill in only the required fields, and the others will be by default.
You may send these parameters:
{
"active": false,
"name": "Test_camera_group",
"comment": "AAAAAA",
"deduplicate": false,
"deduplicateDelay": 0,
"labels": {},
"permissions": {},
"face_threshold": 0,
"body_threshold": 0,
"car_threshold": 0
}
If the response is successful (Created: 201), it returns an object that contains parameters. Example.
Response example
{
"id": 2,
"created_date": "2023-01-23T12:14:09.885006Z",
"modified_date": "2023-01-23T12:14:09.885030Z",
"active": false,
"name": "Test_camera_group",
"comment": "AAAAAA",
"deduplicate": false,
"deduplicateDelay": 0,
"labels": {},
"permissions": {
"1": "edit"
},
"face_threshold": 0,
"body_threshold": 0,
"car_threshold": 0
}
Use camera group id to make a POST request for creating a new camera.
Useful requests
GET /camera-groups/
POST /camera-groups/
GET /camera-groups/{id}/
DELETE /camera-groups/{id}/
PATCH /camera-groups/{id}/
GET /camera-groups/count/
Cameras
Create a new camera
To add a new camera, use this method:
POST /cameras/
The REQUEST BODY contains application/json object with the following required parameters:
Name |
Schema |
Description |
|---|---|---|
|
integer |
Camera group. |
|
string |
Camera name. Constraints: 1 to 256 chars. |
The other parameters you can see at http://<findface-ip:port>/api-docs.
You may send the request with these parameters substituting with your values:
Request example
{
"group": 1,
"active": false,
"name": "Camera_name",
"comment": "AAAAAA",
"url": "url"
}
CURL example
curl -X POST "http://<findface-ip:port>/cameras/" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
-H "Content-Type: application/json" \
-d '{"group":1,"active":false,"name":"Camera_name","comment":"AAAAAA","url":"url"}' \
Returns
Created: 201 – if success.
Search Faces in System
FindFace Multi allows you to search for individuals throughout the entire system.
You may specify a face to search for in one of the following ways:
by uploading a photo
by event ID
by object ID
by cluster ID
To find an individual by uploading a photo, do the following:
Upload a photo with the
POST /detectmethod.Attach source image file and send the
POSTrequest.In the response you will get
idof the object and the coordinates of bbox of the face.If there are multiple faces in the image, select the one of your interest. Copy the value of
idof the necessary object to use it for search withGET /cards/humans/method in thelooks_likefield in this way:detection:<detection id>.You may also use:
event:<event id>(face, body, or car)object:<object id>(face, body or car)cluster:<cluster id>
You will see the search results in response. Example. If necessary, you can narrow down your search by specifying a watch list, similarity threshold, etc.
Compare Two Faces
FindFace Multi allows you to compare two faces and verify that they belong to the same individual.
You may compare two objects from different sources or object with an object within the same card and return the similarity between them.
Do the following:
Perform face (car, body) detection on the attached photo and copy
idof the necessary object.Use the
GET /verifymethod.
GET /verify
The REQUEST contains QUERY-STRING PARAMETERS:
Name |
Schema |
Description |
|---|---|---|
|
string |
Min 1 chars. This field accepts the id of the card that holds object to compare. |
|
string |
Min 1 chars. This field can contain one of the following references:
|
|
string |
Min 1 chars. Same as object1. |
CURL example
curl -X GET "http://<findface-ip:port>/verify?card_id=1&object1=detection%3Acf58e4uv54rotim9jtd0&object2=detection%3Acf58g1mv54rotim9jtdg" \
-H "Accept: application/json" \
-H "Content-Language: ru" \
-H "Accept-Language: ru" \
-H "Authorization: Token <token>" \
If the response is successful (OK: 200), it returns an object that contains confidence.
Name |
Schema |
Description |
|---|---|---|
|
number |
Face similarity score. Constraints: Min 0┃Max 1. |
Response example
{
"confidence": {
"face_objects": {
"4493493039043981648": 0.7896046
},
"average_conf": 0.7896046
}
}
Webhooks
You can set up FindFace Multi to automatically send notifications about specific events, episodes, and counter records to a given URL. To do so, create and configure a webhook. In this case, when such an event, episode, or a counter record occurs, FindFace Multi will send an HTTP request to the URL configured for the webhook.
You can use webhooks for various purposes, for instance, to notify a user about a specific event, invoke required behavior on a target website, and solve security tasks such as automated access control.
In this section:
Configure Webhook
Important
The user must have corresponding permissions to create, view, change or delete a webhook. To manage permissions, navigate to Settings → Role.
Note
To use the webhooks, make sure that at least one of the following parameters is specified in /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py: SERVICE_EXTERNAL_ADDRESS or EXTERNAL_ADDRESS.
To create and configure a webhook, do the following:
Navigate to the Settings tab. Click Webhooks.
Click + New webhook.
Specify the webhook header.
Specify URL to automatically send notifications to.
You can send notifications in batches. Specify the maximum number of notifications in a webhook batch. The actual number may be less.
Specify the maximum number of attempts to send a notification. The interval between attempts increases exponentially with a maximum of 100 seconds.
Important
To receive all messages since the connection loss, should it happen, set
0. Set1to omit old messages.FindFace Multi will be automatically sending notifications on events, episodes, and counter records that match given filters.
There are two ways to configure filters:
visually
by code
To configure filters visually, click Configure.
Generic filters for recognition events (face, body, vehicle):
Matches: filter events only with/without matches, or all events.
Watch lists: filter events only for a selected watch list, or all events.
Camera groups: filter only events from a selected group of cameras.
Cameras: filter only events from a selected camera.
Record name: filter events with a given record name.
Event’s best shot: filter all events of a track, only events with real-time snapshots, only one event with the best snapshot at the end of a track.
Line: filter events with a given line crossing.
Line crossing direction: filter events with line crossing of a given direction, or all events.
Has reverse crossing: filter events with/without reverse crossing.
Filters specific for face recognition events:
Age: filter events with people of a given age.
Beard: filter events by the fact of having a beard.
Emotions: filter events with given emotions.
Gender: filter events with people of a given gender.
Glasses: filter events by the fact of wearing glasses.
Liveness: filter events by face liveness.
Face mask: filter events by the fact of wearing a face mask.
Head turn: filter events by degree of head turn.
Head tilt: filter events by degree of head tilt.
Filters specific for body recognition events:
Gender by body: filter only events with people of a given gender or all events.
Age by body: filter only events with people of a given age.
Headwear: filter only events with a person wearing headgear of a given type: hat/cap, hood/headscarf, none.
Vest: filter only events with a person wearing a vest of a given color.
Helmet: filter only events with a person wearing a helmet of a given color.
Upper clothes color: filter only events with a person wearing a top of a given color.
Lower clothes color: filter only events with a person wearing a bottom of a given color.
Upper clothes type: filter only events with a person wearing upper body wear of a given specific type: jacket, coat, sleeveless vest, sweatshirt, T-shirt, shirt, dress.
Lower body clothes: filter only events with a person wearing lower body wear of a given type: pants, skirt, shorts, obscured.
Upper body clothes: filter only events with a person wearing upper body wear of a given generalized category: long sleeves, short sleeves, no sleeve.
Bag on the back: filter only events with a person wearing/not wearing a bag on the back.
Bag in hand: filter only events with a person wearing/not wearing a bag in hand.
Filters specific for vehicle recognition events:
Make: filter vehicle events by vehicle make.
Model: filter vehicle events by vehicle model.
Vehicle body type: filter only events with vehicles of a given body type: suv, sedan, crossover, convertible, coupe, shooting brake, pickup, van, minivan, limousine.
Vehicle body color: filter only events with vehicles of a given color.
Country: filter only events with vehicles registered in a given country.
License plate number: filter an event with a given plate number.
Special vehicle: filter only events with vehicles belonging to a given type: police, fire service and EMERCOM vehicles, gas rescue and emergency services, military, municipal vehicles, and others.
Vehicle category: filter only events with vehicles belonging to a given category: motorcycle, scooter, car, car with a trailer, truck, truck with a trailer, bus, articulated bus, and others.
Vehicle weight and body size: filter only events with vehicles of a given weight and body size.
Vehicle orientation: filter only events with vehicles of a given orientation.
Note
License plate color, Region are disabled by default. To enable these parameters, uncomment
"license_plate_number_color","license_plate_region"in the sectionFFSECURITY_UI_CONFIGin the configuration file/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py.sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FFSECURITY_UI_CONFIG = { ... "cars": { "features": { ... "license_plate_number_color": [ "unknown", "white", "yellow", "blue", "green", "grey", ], ... "license_plate_region": ["unknown", "DXB", "SHJ", "ADH", "AJM", "UAK", "RAK", "FUJ"], ... }, .... }, }
Restart the
findface-multi-findface-multi-legacy-1container after you have applied the changes.sudo docker container restart findface-multi-findface-multi-legacy-1
Episodes with individuals:
Episode types: filter episodes by status (open/have events/close).
Matches: faces: filter episodes only with/without face matches, or all events.
Matches: bodies: filter episodes only with/without body matches, or all events.
Watch lists: filter episodes only for a selected watch list, or all episodes.
Camera groups: filter only episodes from a selected group of cameras.
Cameras: filter only episodes from a selected camera.
Record name: filter episodes with a given record name.
Count events: filter only episodes with a given number of events.
Episodes with vehicles:
Episode types: filter episodes by status (open/have events/close).
Matches: filter episodes only with/without matches, or all events.
Watch lists: filter episodes only for a selected watch list, or all episodes.
Camera groups: filter only episodes from a selected group of cameras.
Cameras: filter only episodes from a selected camera.
Record name: filter episodes with a given record name.
Episode ID: filter an episode with a given ID.
Count events: filter only episodes with a given number of events.
Counter records:
Camera groups: filter only counter records from a selected group of cameras.
Cameras: filter only counter records from a selected camera.
Count faces ≥: filter counter records by count of faces.
Count bodies ≥: filter counter records by count of bodies.
Count vehicles ≥: filter counter records by count of vehicles.
Counter ID: filter counter records with a given ID.
Minimum distance: send a notification if the detected minimum distance is in specified range.
Maximum distance: send a notification if the detected maximum distance is in specified range.
Average distance: send a notification if the detected average distance is in specified range.
For a number of filters there’s also an option to specify a confidence threshold.
Note
Several parameters are available in the code mode only and are not represented as visual filters. To properly configure these parameters by code, read their descriptions.
To configure filters by code, switch to the Code view. Use the following descriptions:
Generic filters for recognition events (face, body, vehicle):
Sections:
face_events,body_events,car_events."matched_lists": watch list ID, [integer]."camera_groups": camera group ID, [integer]."cameras": camera ID, [integer]."matched_card": matched record ID, [integer]."matched": thematchedstatus. Settrueorfalseif only matched or unmatched events must trigger sending a notification, boolean."confidence_gte": minimum confidence, number. Min 0, Max 1."bs_type": object tracking mode, [string]. Possible values:overall,realtime."line": line crossing ID, [integer]."line_crossing_direction": line crossing direction, [enum]. Possible values:AB,BA."backward_line_crossing": Settrueorfalseif event has or hasn’t reverse crossing, boolean.
Filters specific for face recognition events:
Section:
face_events."headpose_yaw_angle_gte": minimum head turn, integer (degrees). If a head is turned fully to the left, the yaw angle is-180. If it is turned fully to the right, it’s180. The neutral head position is0."headpose_yaw_angle_lte": maximum head turn, integer (degrees). If a head is turned fully to the left, the yaw angle is-180. If it is turned fully to the right, it’s180. The neutral head position is0."headpose_pitch_angle_gte": minimum head tilt, integer (degrees). If a head is tilted forward with the chin touching the chest, it’s-180. If a head is tilted backward facing the sky, it’s180. The neutral head position is0."headpose_pitch_angle_lte": maximum head tilt, integer (degrees). If a head is tilted forward with the chin touching the chest, it’s-180. If a head is tilted backward facing the sky, it’s180. The neutral head position is0."gender": gender, [string]. Possible values:"male","female", [string]."age_lte": maximum age, integer."age_gte": minimum age, integer."glasses": glasses, [string]. Possible values:"none","eye"(eyeglasses),"sun"(sunglasses)."emotions": emotions, [string]. Possible values:"any","angry","fear","sad","neutral","disgust","happy","surprise"."beard": beard, [string]. Possible values:"none","beard"."medmask": face mask, [string]. Possible values:"correct","incorrect","none"."liveness": liveness, [string]. Possible values:"real","fake"."temperature_gte": minimum temperature, number."liveness_gte": minimum algorithm confidence that a face belongs to a live person, number. Min 0, Max 1.
Filters specific for body recognition events:
Section:
body_events."headwear": type of headgear (hat/cap, hood/headscarf, none), [string]. Possible values:"hat","hood","none"."upper_clothes": generalized category of upper body wear (long sleeves, short sleeves, no sleeve), [string]. Possible values:"long_sleeves","short_sleeves","without_sleeves"."detailed_upper_clothes": specific type of upper body wear (jacket, coat, sleeveless vest, sweatshirt, T-shirt, shirt, dress), [string]. Possible values:"jacket","coat","sleeveless","sweatshirt","t-shirt","shirt","dress"."top_color": top clothing color, [string]. Possible values:white,"black","grey","brown","red","orange","yellow","green","lightblue","blue","purple","pink","beige","violet"."lower_clothes": type of lower body wear (pants, skirt, shorts, nondescript), [string]. Possible values:"pants","obscured","skirt","shorts"."bottom_color": bottom clothing color, [string]. Possible values:white,"black","grey","brown","red","orange","yellow","green","lightblue","blue","purple","pink","beige","violet"."bag_hand": whether a person has a bag in hand(s), [string]. Possible values:"none","hand"."bag_ground": whether a person has a bag on the ground. Possible values:"ground","none"."bag_back": whether a person has a bag on the back, [string]. Possible values:"none","back"."vest_type_score_gte": minimum recognition confidence of a vest presence, number. Min 0, Max 1."vest_type_score_lte": maximum recognition confidence of a vest presence, number. Min 0, Max 1."vest_type": presence of personal protective equipment in the form of a vest, [string]. Possible values:"green/yellow","orange","not_visible","none"."helmet_type_score_lte": maximum recognition confidence of a helmet presence, number. Min 0, Max 1."helmet_type_score_gte": minimum recognition confidence of a helmet presence, number. Min 0, Max 1."helmet_type": presence of personal protective equipment in the form of a helmet, [string]. Possible values:"red/orange","white","other","not_visible","none"."age_group_score_lte": maximum age group recognition confidence, number. Min 0, Max 1."age_group_score_gte": minimum age group recognition confidence, number. Min 0, Max 1."age_group": age by group, [string]. Possible values:"0-16","17-35","36-50","50+"."gender_score_lte": maximum gender recognition confidence, number."gender_score_gte": minimum gender recognition confidence, number."body_gender": gender, [string]. Possible values:"male","female".
Filters specific for vehicle recognition events:
Section:
car_events."category_confidence_gte": minimum recognition confidence of the vehicle category, number. Min 0, Max 1."category_confidence_lte": maximum recognition confidence of the vehicle category, number. Min 0, Max 1."category_type": vehicle category, [string]. Possible values:"unknown","A","B","BE","C","CE","D","DE","other"."color"vehicle color, [string]. Possible values:"beige","black","blue","brown","cyan","gold","green","grey","orange","pink","purple","red","silver","violet","white","yellow"."body": vehicle body style, [string]. Possible values:"suv","sedan","crossover","convertible","coupe","wagon","cab","minibus","minivan","limousine"."make": vehicle make, [string]. Check out the corresponding event filter to see what makes are supported in the current version."model": vehicle model, [string]. Check out the corresponding event filter to see what models are supported in the current version."license_plate_number": license plate number, [string]. Wildcards are not supported."license_plate_number_color": license plate color, [string]. Possible values:"unknown","white","yellow","blue","green","grey"."license_plate_number_color_confidence_lte": maximum recognition confidence of the license plate color, number."license_plate_number_color_confidence_gte": minimum recognition confidence of the license plate color, number."license_plate_country": license plate country, [string]. Check out the corresponding event filter to see what countries are supported in the current version."license_plate_region": license plate region, [string]. Check out the corresponding event filter to see what regions are supported in the current version."special_vehicle_type": special vehicle type, [string]. Possible values:"taxi","route_transport","car_sharing","ambulance","police","rescue_service","gas_service","military","road_service","other_special","not_special"."weight_type": vehicle weight type, [string]. Possible values:"B_light","B_heavy","C_light","C_heavy","D_light","D_long","other"."weight_type_confidence_lte": maximum recognition confidence of the vehicle weight type, number. Min 0, Max 1."weight_type_confidence_gte": minimum recognition confidence of the vehicle weight type, number. Min 0, Max 1."orientation": vehicle orientation, [string]. Possible values:"front","back","side"."orientation_confidence_lte": maximum recognition confidence of the vehicle orientation, number. Min 0, Max 1."orientation_confidence_gte": minimum recognition confidence of the vehicle orientation, number. Min 0, Max 1.
Episodes with individuals:
Section:
human_episodes."allowed_types": episode status, [string]. Possible values: an episode opening (episode_open), adding a new event into an episode (episode_event), an episode closing (episode_close)."matched_lists": watch list ID, [integer]."cameras": camera ID, [integer]."camera_groups": camera group ID, [integer]."events_count_gte": minimum number of events in an episode, integer."events_count_lte": maximum number of events in an episode, integer."face_matched": thematchedstatus of face events in an episode, boolean. Settrueif only matched faces orfalseif only unmatched faces must trigger sending a notification."body_matched": thematchedstatus of body events in an episode, boolean. Settrueif only matched bodies orfalseif only unmatched bodies must trigger sending a notification.
Episodes with vehicles:
Section:
car_episodes."allowed_types": episode status, [string]. Possible values: an episode opening (episode_open), adding a new event into an episode (episode_event), an episode closing (episode_close)."matched_lists": watch list ID, [integer]."cameras": camera ID, number."camera_groups": camera group ID, [integer]."events_count_gte": minimum number of events in an episode, integer."events_count_lte": maximum number of events in an episode, integer."car_matched": thematchedstatus of car events in an episode, boolean. Settrueif only matched cars orfalseif only unmatched cars must trigger sending a notification.
Counter records:
Section:
counters."counter": counter ID, [integer]."cameras": camera ID, [integer]."camera_groups": camera group ID, [integer]."faces_count_gte": minimum number of faces in a counter record, integer."faces_count_lte": maximum number of faces in a counter record, integer."silhouettes_count_gte": minimum number of bodies in a counter record, integer."silhouettes_count_lte": maximum number of bodies in a counter record, integer."cars_count_gte": minimum number of vehicles in a counter record, integer."cars_count_lte": maximum number of vehicles in a counter record, integer."proximity_min_lte": send a notification if the minimum detected distance in meters is less than this value, number."proximity_min_gte": send a notification if the minimum detected distance in meters is greater than this value, number."proximity_avg_lte": send a notification if the average detected distance in meters is less than this value, number."proximity_avg_gte": send a notification if the average detected distance in meters is greater than this value, number."proximity_max_lte": send a notification if the maximum detected distance in meters is less than this value, number."proximity_max_gte": send a notification if the maximum detected distance in meters is greater than this value, number.
Example
Note
This example is given for reference only, substitute your values in corresponding fields.
{ "name": "AA", "active": false, "url": "http://example.com", "batch_size": 1, "filters": { "face_events": { "matched_lists": [ -1 ], "camera_groups": [ -1 ], "cameras": [ 1 ], "matched_card": [ 1 ], "matched": false, "confidence_gte": 0, "bs_type": [ "A" ], "line": [ 1 ], "line_crossing_direction": [ "AB" ], "backward_line_crossing": false, "headpose_yaw_angle_gte": -180, "headpose_yaw_angle_lte": -180, "headpose_pitch_angle_gte": -180, "headpose_pitch_angle_lte": -180, "gender": [ "A" ], "age_lte": 0, "age_gte": 0, "glasses": [ "A" ], "emotions": [ "A" ], "beard": [ "A" ], "medmask": [ "A" ], "liveness": [ "A" ], "temperature_gte": 0, "liveness_gte": 0 }, "body_events": { "matched_lists": [ -1 ], "camera_groups": [ -1 ], "cameras": [ 1 ], "matched_card": [ 1 ], "matched": false, "confidence_gte": 0, "bs_type": [ "A" ], "line": [ 1 ], "line_crossing_direction": [ "AB" ], "backward_line_crossing": false, "headwear": [ "A" ], "upper_clothes": [ "A" ], "detailed_upper_clothes": [ "A" ], "top_color": [ "A" ], "lower_clothes": [ "A" ], "bottom_color": [ "A" ], "bag_hand": [ "A" ], "bag_ground": [ "A" ], "bag_back": [ "A" ], "vest_type_score_gte": 0, "vest_type_score_lte": 0, "vest_type": [ "A" ], "helmet_type_score_lte": 0, "helmet_type_score_gte": 0, "helmet_type": [ "A" ], "age_group_score_lte": 0, "age_group_score_gte": 0, "age_group": [ "A" ], "gender_score_lte": 0, "gender_score_gte": 0, "body_gender": [ "A" ] }, "car_events": { "matched_lists": [ -1 ], "camera_groups": [ -1 ], "cameras": [ 1 ], "matched_card": [ 1 ], "matched": false, "confidence_gte": 0, "bs_type": [ "A" ], "line": [ 1 ], "line_crossing_direction": [ "AB" ], "backward_line_crossing": false, "category_confidence_gte": 0, "category_confidence_lte": 0, "category_type": [ "A" ], "color": [ "A" ], "body": [ "A" ], "make": [ "A" ], "model": [ "A" ], "license_plate_number": [ "A" ], "license_plate_number_color": [ "A" ], "license_plate_number_color_confidence_lte": 0, "license_plate_number_color_confidence_gte": 0, "license_plate_country": [ "A" ], "license_plate_region": [ "A" ], "special_vehicle_type": [ "A" ], "weight_type": [ "A" ], "weight_type_confidence_lte": 0, "weight_type_confidence_gte": 0, "orientation": [ "A" ], "orientation_confidence_lte": 0, "orientation_confidence_gte": 0 }, "human_episodes": { "allowed_types": [ "A" ], "matched_lists": [ -1 ], "cameras": [ 1 ], "camera_groups": [ -1 ], "events_count_gte": 0, "events_count_lte": 1, "face_matched": false, "body_matched": false }, "car_episodes": { "allowed_types": [ "A" ], "matched_lists": [ -1 ], "cameras": [ 1 ], "camera_groups": [ -1 ], "events_count_gte": 0, "events_count_lte": 1, "car_matched": false }, "counters": { "counter": [ 0 ], "cameras": [ 0 ], "camera_groups": [ -1 ], "faces_count_gte": 1, "faces_count_lte": 0, "silhouettes_count_gte": 1, "silhouettes_count_lte": 0, "cars_count_gte": 1, "cars_count_lte": 0, "proximity_min_lte": 0, "proximity_min_gte": 0, "proximity_avg_lte": 0, "proximity_avg_gte": 0, "proximity_max_lte": 0, "proximity_max_gte": 0 }, "areas": { "area_in": [ 0 ], "camera_in": [ 0 ], "camera_group_in": [ -1 ] } }, "send_attempts": 0 }
Important
Use only filters which match your search needs. To turn off a filter, remove it from a webhook. Do not leave a filter empty (
[]) as in this case, the result of filtration will be empty as well.Note
To get all notifications, pass only curly braces without any enclosed filters:
{}Tip
Example #1. Get notifications about all vehicle events:
{ "car_events": {} }
Example #2. Get notifications of the opening of matched episodes that feature human faces and bodies:
{ "human_episodes": { "allowed_types": ["episode_open"], "face_matched": true, "body_matched": true }}
Note
You can specify several values for filters with square braces. In this case, the webhook will be triggered once one of the values from this filter has been matched. In the example below, you will get a body event from the camera group
1or3if a matched record is12or25.{ "body_events": { "camera_groups": [1, 3], "matched_card": [12, 25], }, }
Check Active.
Click Save.
Webhook in Action
Try out a webhook by capturing event notifications with a simple web server in Python:
from pprint import pprint
from aiohttp import web
async def handle(request):
pprint(await request.json())
return web.Response(status=200)
app = web.Application()
# for aiohttp v 3.x
# app.add_routes([web.post('/', handle)])
# for aiohttp v 2.x
app.router.add_post('/', handle)
web.run_app(app, port=8888)
Important
A webhook catcher that you use must return an HTTP 200 response after receiving the webhook request from FindFace Multi, like in the example above.
If no filters are configured for a webhook, this web server will be getting notifications about all events, episodes, and counter records.
To view a webhook pulling status, execute:
sudo journalctl CONTAINER_NAME=findface-multi-findface-multi-legacy-1 -f | grep 'Webhook'
Verbose Webhooks
By default, webhook notifications contain only IDs of such entities as cards, watch lists, cameras, and camera groups. It is possible to get the full content of the mentioned entities by switching webhooks to the verbose mode.
To do so, open the /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py configuration file and set 'VERBOSE_WEBHOOKS': True:
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py
...
FFSECURITY = {
...
# send serialized cards, card-lists, camera and camera groups in webhooks
'VERBOSE_WEBHOOKS': True,
...
}
...
Be sure to restart the findface-multi-findface-multi-legacy-1 container after making changes.
sudo docker container restart findface-multi-findface-multi-legacy-1
External VMS
On the FindFace Multi side, a mechanism of integration with third-party VMS is implemented, allowing you to add cameras from the latter to FindFace Multi and perform their processing. When a camera from a VMS is added, all the actions available for a regular FindFace Multi camera become available for it.
Integration with third-party VMS is possible through a plug-in layer. Interaction between the plugin and FindFace Multi is performed via HTTP. The plugin interacts with the VMS, FindFace Multi interacts with the plugin.
In this section:
This is the step-by-step configuration guide for installing and configuring External VMS feature in FindFace Multi. The reader should have strong experience of administering Linux OS and Docker containers.
Create a Service Account for VMS Integration Plugin
Create a user (username: ntech) with specific set of permissions. This will be used as the service account for the integration plugin. To do so, execute the command below:
adduser --system --disabled-password --disabled-login --home /var/empty \
--no-create-home --quiet --force-badname --group ntech
Configure FindFace Multi to Enable External VMS Integration Plugin
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyfile and locate theExternal VMS Integration settingssection.Change the
'ENABLED'parameter toTrue.Write down the
TOKENvalue, as you will need it to configure the plugin later. The default value isPLUGIN_TOKEN. Please note that theVmsPluginTokentoken prefix should only be in thefindface-multi-legacy.pyconfiguration file.... # -- External VMS integration settings -- # cleanup settings 'EXTERNAL_VMS_EVENTS_MAX_AGE': 0, 'EXTERNAL_VMS_SEND_EVENTS_STATUS_MAX_AGE': 0, 'EXTERNAL_VMS': { 'ENABLED': True, 'PLUGIN_ADDRESS': 'http://127.0.0.1:18333', 'TOKEN': 'VmsPluginToken PLUGIN_TOKEN', 'EVENT_SENDER': { 'ENABLED': True, 'ALLOWED_TYPES': ['face'], 'SENDER_TASKS': 1, 'MAX_SEND_ATTEMPTS': 1, 'MIN_EVENT_SEND_TIMEOUT': 0.1, 'MAX_EVENT_SEND_TIMEOUT': 100, 'RESPONSE_TIMEOUT': 10, }, }, ...
Tip
In our example, we are installing the plugin on the same host where FindFace Multi installation resides. In case of using the separate machine to install the plugin, please consider specifying the proper address of the server where the VMS integration plugin resides in the
PLUGIN_ADDRESSparameter.Note
If you need to send the events to VMS, please also set the
'ENABLED'value toTruein theEVENT_SENDERsection.Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
Brief description of the parameters available below:
Parameter |
Description |
|---|---|
EXTERNAL_VMS_EVENTS_MAX_AGE |
Time to store the events received from an external VMS system (days). 0 – unlimited. |
EXTERNAL_VMS_SEND_EVENTS_STATUS_AGE |
Time to store the events sent to an external VMS system (days). 0 – unlimited. |
EXTERNAL_VMS: ENABLED |
The status of FindFace VMS Integration module. If enabled, the |
PLUGIN_ADDRESS |
Plugin IP address. |
TOKEN |
Token required for VMS Integration plugin to authenticate in Findface Multi (it must match the token in the plugin configuration file). |
EVENT_SENDER:ENABLED |
If |
ALLOWED_TYPES |
Objects types to send. At present, only face objects supported. |
SENDER_TASKS |
The number of simultaneous tasks to send events to the plugin. Default value = 1. It can be increased if required, please consult NtechLab Support Team (support@ntechlab.com). |
MAX_SEND_ATTEMPTS |
The number of attempts to send before it is considered unsuccessful. |
MIN_EVENT_SEND_TIMEOUT MAX_EVENT_SEND_TIMEOUT |
The time between sending attempts, if a failed sending occurs. It will grow exponentially from minimum to maximum as long as the number of attempts MAX_SEND_ATTEMPTS is increasing. |
RESPONSE_TIMEOUT |
Time to wait for a response from the plugin when sending an event. |
Create New Database in PostgreSQL
Open the file
/opt/findface-multi/docker-compose.yamland write down thePOSTGRES_PASSWORDvalue. You will need it for the next step.sudo vi /opt/findface-multi/docker-compose.yaml ... postgresql: environment: {POSTGRESQL_ALLOW_REMOTE_CONNECTIONS: 'no', POSTGRES_PASSWORD: POSTGRES_PASSWORD} ...
Enter the
findface-multi-postgresql-1container via executing the following command:sudo docker exec -i findface-multi-postgresql-1 /bin/bash -c "PGPASSWORD={POSTGRES_PASSWORD} psql --username postgres"
To ensure you’ve signed in successfully, use the
\lcommand to view the list of databases:
Run the following command to create the new database required for external VMS functionality:
CREATE DATABASE ffsintegration WITH OWNER ntech ;
Validate the database creation via executing
\lagain. Once completed, use\qparameter to exit PostgreSQL.
Install and Configure FindFace VMS Integration Plugin
Open the
/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyfile and locate theDATABASESsection. Write down thePASSWORDvalue for userntech:sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py ... # camera groups, watchlists and so on. Only PostgreSQL is supported. DATABASES = { 'default': { 'ENGINE': 'django.db.backends.postgresql', 'DISABLE_SERVER_SIDE_CURSORS': True, 'NAME': 'ffsecurity', 'HOST': '127.0.0.1', 'PORT': 5439, 'USER': 'ntech', 'PASSWORD': 'PASSWORD' } }
Request the plugin
findface-vms-integration-plugin.imgfrom an NtechLab representative.Load the plugin into the system.
docker load < findface-vms-integration-plugin.img
Configure the
/opt/findface-multi/docker-compose.yamlfile to add thefindface-vms-integration-pluginsection.cd /opt/findface-multi/ sudo vi docker-compose.yaml
Add the
findface-vms-integration-pluginsection to the end of thedocker-compose.yamlfile.findface-vms-integration-plugin: depends_on: [findface-ntls, postgresql, findface-multi-legacy] image: findface-vms-integration-plugin:latest network_mode: service:pause volumes: ['./configs/findface-vms-integration-plugin/findface-vms-integration-plugin.conf:/etc/findface-vms-integration-plugin.conf'] restart: always
Create the configuration file for the plugin.
sudo mkdir configs/findface-vms-integration-plugin/ sudo vi configs/findface-vms-integration-plugin/findface-vms-integration-plugin.conf
Below is a sample of the
findface-vms-integration-plugin.confconfiguration file.# URL prefix for API version 1 endpoints # Default: /api/v1 # API_V1_STR: /v1 # App listens on this interface. Set it to 0.0.0.0 to listen on all available interfaces # Default: 127.0.0.1 # # HOST: 127.0.0.1 # App accepts HTTP connections on this port # Default: 18333 # # PORT: 18333 # Public URL for App # Default: http://127.0.0.1:18333 # # PUBLIC_URL: http://127.0.0.1:18333 # Public URL for FFS # Default: http://127.0.0.1 # # FFS_PUBLIC_URL: http://127.0.0.1:18333 # Full database URL dialect+driver://username:password@host:port/database. # Alternatively, connection parameters can be specified one by one (see below). # # SQLALCHEMY_DATABASE_URI: # Database host # POSTGRES_SERVER: 127.0.0.1 # Database port number # POSTGRES_PORT: 5432 # Database user # POSTGRES_USER: ntech # Database user password # POSTGRES_PASSWORD: POSTGRES_PASSWORD # Database name # POSTGRES_DB: ffsintegration # Log SQL statements # Default: FALSE # # LOG_QUERIES: FALSE # Secret token used to validate requests made by FindFace Security # Required: yes # FFS_TOKEN: PLUGIN_TOKEN # Include traceback information in error responses # Default: FALSE # RESP_TRACEBACK: True # Logging level # Default: INFO # # LOG_LEVEL: # Enable journald log handler. # Default: TRUE # # LOG_TO_JOURNALD: TRUE # Timeout for requests to VMS # Default: 10 # # VMS_REQUEST_TIMEOUT_SEC: 10 # Timeout for requests to FindFace # Default: 10 # # FF_REQUEST_TIMEOUT_SEC: 10
Tip
In our example, we are installing the plugin on the same host where FindFace Multi installation resides. In case of using a separate machine, please consider specifying the proper address of FindFace Multi server in the
POSTGRES_SERVERparameter and the FindFace Multi URL in theFFS_PUBLIC_URLparameter.If you installed the plugin on the FindFace Multi host, you only need to configure the login and password to the local PostgreSQL database.
POSTGRES_USER: ntech POSTGRES_PASSWORD: POSTGRES_PASSWORD
The
POSTGRES_PASSWORDvalue can be obtained from the/opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.pyfile.To enable the external VMS plugin in FindFace Multi start all FindFace Multi containers from the
/opt/findface-multidirectory.sudo docker-compose up -d
Open FindFace Multi in a web browser. You will see the Settings -> External VMS section.
Warning
It is necessary to set the same time zone on the external VMS server as on the FindFace Multi server for correct playback of video archives from the external VMS.
Add External VMS in FindFace Multi
In the Settings -> External VMS section, click on the Add new VMS button.
Specify your VMS parameters in the opened windows and click on the Save button once ready:
Once completed now you can manage the streams from your VMS in the Cameras tab. For instance, you may add all or selected cameras in FindFace Multi instantly.
Should you need to send the events to an external VMS, in the Camera settings added from VMS, check if it configured to send the events to the external VMS:
In the required watch list properties, activate the checkbox
Send mapping events to VMSand click on Save button to apply the changes.
Sent and received messages can be viewed in the corresponding tabs of the target VMS:
If a Video Recorder is deployed and configured, you can play the video archives of the external VMS (Axxon, Trassir and Hikcentral) through it. To use this functionality, first check the camera added from VMS is active. Then navigate to the Episodes & Events tab and hover over the event or episode of your interest. The play button will display on the thumbnail of the object (face/body or vehicle). Press it to start playing the video from the time when the event/episode was recorded from the camera.
You can also play the video archive through the camera by clicking play on the camera thumbnail in the camera list on the Video Sources tab. As a result, the video from the external VMS camera will be displayed in the Video Recorder. If there is a recorded video archive in FindFace Multi at the specified time, then it will be played in the Video Recorder instead of the video from an external VMS.
Now, your FindFace Multi installation is enhanced by external VMS support. You may connect as many VMS systems and required and add existing video streams in FindFace seamlessly within just a few clicks.
External Detectors
It is possible to integrate FindFace Multi with the external detectors that can source frames for object recognition, e.g., Access Control terminals. In this case, once FindFace Multi receives a frame from an external detector, it will initiate the extraction of the object’s feature vector and event creation. You can work with these events by analogy with the events from CCTV cameras.
Important
To use an external detector for object recognition, liveness detection must be enabled. If you have not enabled it during installation, you can do it now, using information from the Enable Face Liveness Detection section.
The liveness model liveness.web.v0 comes in the default configuration and is used for authentication by face. If your use case implies connection to an external detector for object recognition, replace the liveness.web.v0 model with the liveness.pacs.v2 model in the findface-extraction-api.yaml configuration file. Since both models employ the same normalizer facenorm.multicrop_full_crop2x_size400.cpu.fnk there is no need to change it.
Replace the liveness model in the /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml configuration file:
GPU
sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
extractors:
...
models:
face_liveness: faceattr/liveness.pacs.v2.gpu.fnk
...
CPU
sudo vi /opt/findface-multi/configs/findface-extraction-api/findface-extraction-api.yaml
extractors:
...
models:
face_liveness: faceattr/liveness.pacs.v2.cpu.fnk
...
Restart the findface-multi-findface-extraction-api-1 container.
sudo docker container restart findface-multi-findface-extraction-api-1
The external detector integration is done through HTTP API. After the primary configuration, FindFace Multi will issue a token. Specify this token in every API request sent by the external detector to FindFace Multi to authorize the device.
To integrate an external detector with FindFace Multi, do the following:
Navigate to the Video Sources -> External Detectors.
Click + Add or Add external detectors if you don’t have any detectors yet.
In the opened tab, specify the external detector name and description. Add the external detector to a camera group, so it will be more convenient to filter events from this device later.
Tip
You can allot a separate camera group specifically to external detectors.
On the Add and configure tab, enable liveness. On the same tab you will see a token.
Specify this token in every API request that the external detector sends to FindFace Multi to create an event. As a result, frames passed in the requests will be associated with the external detector’s relevant camera and processed by analogy with the frames from CCTV cameras.
Detailed interactive documentation on the FindFace Multi HTTP API is available after installation at
http://<findface-ip:port>/api-docs. Learn and try it out.Tip
You can also find it by navigating to Settings -> API Documentation in the web interface.
Alarm Monitor
In this chapter:
FindFace Multi allows for embedding an alarm monitor service, an additional functionality that displays information about individual alarms and events as they occur. This increases the awareness of the security service regarding critical events. An alarm monitor cannot be blocked by other windows, and the configuration cannot be changed to mute, which makes it a simpler tool for monitoring centers. Operators can acknowledge alarms or discard.
Server Preparation
For working alarm monitor service, FindFace Multi uses Annex, a software platform that provides video analytics tools. Annex is installed by default when you deploy FindFace Multi, and includes a set of components and services:
Component |
Ports in use |
Description |
Vendor |
|---|---|---|---|
backend_provider |
1111 |
Service that provides an interface (REST API) for delivering data to various microservices, enriches input data if necessary and converts them to the desired format. |
NtechLab own deployment |
backend_api |
2222 |
Service that provides an interface (REST API) to application service databases, as well as a proxy interface to FindFace Multi and FindFace Server. |
|
service_notifier |
3311 |
The auxiliary service for sending notifications (Notify) through various client channels (WebSocket, Telegram, Mail), as well as for sending client responses (Reply) back to the services. |
|
alarm-app |
4444 |
A web service of the Alarm Monitor. |
|
service_alarmer |
n/a |
An application service for announcing alarm messages with the ability to acknowledge them by the user. |
Annex receives events from FindFace Multi via webhooks and sends them to Alarm Monitor.
Before runnning Alarm Monitor, make sure Annex is deployed and perform some preliminary actions in FindFace Multi.
Make sure all Annex services are running.
sudo docker ps -a
In order Alarm Monitor to work correctly with Annex, it is necessary to enable the VERBOSE_WEBHOOKS parameter to get information about the webhook, line crossing, matched cards as well as SAVE_DETECTOR_PARAMS to get information about the bbox.
sudo vi /opt/findface-multi/configs/findface-multi-legacy/findface-multi-legacy.py FFSECURITY = { ... # send serialized cards, card-lists, camera and camera groups in webhooks 'VERBOSE_WEBHOOKS': True, ... # save full detector params from video worker request 'SAVE_DETECTOR_PARAMS': True, ... }
Restart the
findface-multi-findface-multi-legacy-1container.sudo docker container restart findface-multi-findface-multi-legacy-1
In the web interface FindFace Multi create a user. See Role and User Management. Use these credentials to log in Alarm Monitor.
Setting Events
In order for the alarm monitor service to be able to receive events of matching objects, line crossing and counter events from FindFace Multi, use webhooks.
To monitor events, do the following:
In the FindFace Multi web interface, navigate to the Settings tab. Click Webhooks.
Click Add new webhook.
Specify the webhook header.
Specify URL to automatically send notifications
http://ip_address:port/ff_multi/alarmer/.Note
ip_address:portmust be substituted into your IP address for Annex and a port forbackend_provider, for examplehttp://127.0.0.1:1111/ff_multi/alarmer/.Specify filters.
Setting Application
The application is available for different platforms and operating systems.
The application is deployed in the standard way for your operating system and also available from the FindFace Multi navigation bar in the web-interface.
Note
We recommend installing Alarm Monitoring Application. Please contact our support and sales team on support@ntechlab.com to get the latest version.
Do the following:
Download the latest version of Alarm Monitoring Application for your operating system.
Run the executable file to launch the application.
Navigate to the settings.
Note
If you use a web interface for running the Alarm Monitor, enter
http://ip_address:4444/config/in the address bar to navigate to the settings. Whereip_addressis the IP address available for Annex.Select the alarm sound and volume.
If necessary, test the alarm sound by clicking the button Imitation Alarm.
Specify the web service address, for example
ws://ip_address/annex.Note
ip_addressis the IP address available for Annex.Select the interface language.
Click Save.
Click Log in.
Enter username and password.
Work with the Alarm Monitor
Alarm Monitor does not display any information by default. When receiving the alarm, the application displays information about the event together with an audio signal. The application will alternately display events until each of them is confirmed.
The following types of alarm events are supported by Alarm Monitor:
Line crossing (the line and track are currently displayed only for bodies and faces)
Matches/non-matches with lists (for faces, bodies and vehicle)
Counters (bbox are currently displayed only for bodies)
Example of line crossing event
The following elements are displayed:
Full frame of the event with the bbox of the body, the line crossing and the track of the object
Photo of the body crossing the line
Webhook header
Camera group/Camera
Date and time of the event
Example of object (face/body) matching event
The following elements are displayed:
Full frame of the event
Photo of the object from the video stream
Photo of the matched object from the record index and a match probability
Watch lists that contain a person/vehicle
Webhook header
Camera group/Camera
Date and time of the event
Example of counter event
The following elements are displayed:
Full frame of the event
Screenshot
Counter name and number of bodies in the screenshot
Webhook header
Camera group/Camera
Date and time of the event
Click Accept to acknowledge the alarm, or No to discard. Alarm confirmation is expected within 90 seconds by default. After the interval expires, an automatic confirmation occurs. You may change this value using the ack_interval parameter in the webhook URL, for example http://127.0.0.1:1111/ff_multi/alarmer/?ack_interval=180. If the value is 0, no confirmation is expected.
Analytical BI Service
Analytical BI service is a service for collecting, storing, visualizing, and interpreting dataset for some period according to a certain logic. The service works in conjunction with FindFace Multi in the following way:
FindFace Multi generates dataset.
Analytical service collects, consolidates, and stores data.
BI system serves for data visualization.
At the moment analytics works for line crossing only.
Analytical Service and Superset BI Deployment
The analytical service and the Superset BI system are deployed together with FindFace Multi from one distribution package. Select the appropriate items during installation.
Q: Do you want to install Analytics service?(y/n)
A: Answer y to install the analytical service.
Q: Do you want to install Superset BI?(y/n)
A: By answering y you choose to install the Apache Superset open-source software application. If you prefer your own BI system, you can skip the suggested installation and answer n.
Preliminary Actions in FindFace Multi
In order for the data to appear in the BI system, perform some preliminary actions in FindFace Multi. Define a virtual polygonal line on a certain camera for an object within FindFace Multi. Detailed information on creating a line is provided in the Line Crossing section. In the BI Analytics tab, add information about the line that you have created. Setting the line attributes will help you to combine different data in one graph in the BI system or sort out unnecessary data.
After you have created a line, the system will generate line crossing events indicating the direction of crossing, and you will be able to see the number of line crossings (traffic) filtered by the direction of the crossing in a graph. You will also be able to compare the same line crossing data over multiple periods.
Work with the Superset BI system
Once you have installed the analytical service and the Superset BI system, you will see the BI Analytics button in the web interface navigation bar. Pressing the BI Analytics button will open a third-party BI system in a separate tab. The BI system web interface address will depend on the interface that you have selected during FindFace Multi installation in step #4.5, the port in use is 8088, e.g., 192.168.112.254:8088. The default credentials are admin | admin. A password can further be reset in the Settings -> Info tab (here and on the information is provided for Apache Superset BI system).
In the FF.Analytics dashboard view overall traffic or object-specific traffic. The filters on the left-side panel may come in handy.
In the object-specific view, monitor traffic attributes (for a face and a silhouette: gender, age, presence of a mask or PPE, and other attributes; for a vehicle: a category, a vehicle make, belonging to a special vehicle, and other attributes).
To enable the required language in the UI, uncomment it by removing # in the configuration file /opt/findface-multi/configs/superset/superset_config.py.
sudo vi /opt/findface-multi/configs/superset/superset_config.py LANGUAGES = { "en": {"flag": "us", "name": "English"}, # "es": {"flag": "es", "name": "Spanish"}, # "it": {"flag": "it", "name": "Italian"}, # "fr": {"flag": "fr", "name": "French"}, # "zh": {"flag": "cn", "name": "Chinese"}, # "ja": {"flag": "jp", "name": "Japanese"}, # "de": {"flag": "de", "name": "German"}, # "pt": {"flag": "pt", "name": "Portuguese"}, # "pt_BR": {"flag": "br", "name": "Brazilian Portuguese"}, "ru": {"flag": "ru", "name": "Russian"}, # "ko": {"flag": "kr", "name": "Korean"}, # "sk": {"flag": "sk", "name": "Slovak"}, # "sl": {"flag": "si", "name": "Slovenian"}, # "nl": {"flag": "nl", "name": "Dutch"}, }
Restart the findface-multi-superset-1 container.
sudo docker container restart findface-multi-superset-1
On the top bar, select your language in the drop-down menu:
