extraction-api

The extraction-api service 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).

It interfaces with the sf-api service as follows:

• Gets original images with objects and normalized object images.

• Returns the coordinates of the object bounding box, and (if requested by sf-api) feature vector and object attribute data.

Tip

You can use HTTP API to directly access extraction-api.

Functionality:

• object detection in an original image (with a return of the bbox coordinates),

• object normalization,

• feature vector extraction from a normalized image,

• object attribute recognition (gender, age, emotions, vehicle model, vehicle color, etc.)

The extraction-api service can be based on CPU (installed from the docker.int.ntl/ntech/universe/extraction-api-cpu image) or GPU (installed from the docker.int.ntl/ntech/universe/extraction-api-gpu image). For both CPU- and GPU-accelerated services, configuration is done through environment variables and command line flags. You can also use the extraction-api configuration file. Its content varies subject to the acceleration type. You can find its default content here for CPU and here for GPU or by running docker command with a flag -config-template.

When configuring extraction-api (on CPU or GPU), refer to the following parameters:

Command line flags	Type	Description
-allow-cors	–	Add CORS headers to allow cross-origin requests.
-ascend-device	uint	Ascend device ID on which to launch inference.
-cache-dir	string	Directory for GPU model’s cache (default /var/cache/findface/models_cache).
-config	string	Path to config file.
-config-template	–	Output config template and exit.
-debug	–	Enable verbose logging.
-detectors-instances	int	DEPRECATED [use detectors-max-batch-size] Number of parallel detector instances.
-detectors-max-batch-size	int	Upper limit on detection batch size. When using the CPU, you can specify max_batch_size: -1, then the detector maximum batch size will correspond to the number of CPU cores (default 1).
-extractors-instances	int	DEPRECATED [use extractors-max-batch-size] Number of parallel extractor instances.
-extractors-max-batch-size	int	Upper limit on extraction batch size. When using the CPU, you can specify max_batch_size: -1, then the extractor maximum batch size will correspond to the number of CPU cores (default 1).
-extractors-models	value	Attribute models.
-fetch-enabled	–	Enable fetching from remote urls (default true).
-fetch-size-limit	int	File size limit (default 10485760).
-gpu-device	uint	GPU device ID on which to launch inference.
-help	–	Print help information.
-license-ntls-server	string	NTLS (ntechlab license server) address (default 127.0.0.1:3133).
-listen	string	IP:port to listen on (default :18666).
-max-dimension	int	Maximum dimension (default 6000).
-models-root	string	Root directory for model files (.fnk) (default /usr/share/findface-data/models).
-normalizers-instances	int	DEPRECATED [use normalizers-max-batch-size] Number of parallel normalizer instances.
-normalizers-max-batch-size	int	Upper limit on normalization batch size. When using the CPU, you can specify max_batch_size: -1, then the normalizer maximum batch size will correspond to the number of CPU cores (default 1).
-normalizers-models	value	Map of facenkit normalizers and their parameters.
-ticker-interval	int	Interval between ticker lines in log in milliseconds (0 - disabled) (default 5000).

The format of environment variable for flag -my-flag is CFG_MY_FLAG.

Priority:

1. Defaults from source code (lowest priority).

2. Configuration file.

3. Environment variables.

4. Command line.

If necessary, you can also enable recognition models for face attributes, body and body attributes, vehicle and vehicle attributes, and liveness detection. You can find the detailed step-by-step instructions in the following sections:

• Enable Face and Face Attribute Recognition.

• Enable Face Liveness Detection

• Liveness Detection as Standalone Service

• Enable Body and Body Attribute Recognition

• Enable Vehicle and Vehicle Attribute Recognition

Important

The acceleration type for each model must match the acceleration type of extraction-api: CPU or GPU. Note that extraction-api on CPU can work only with CPU-models, while extraction-api on GPU supports both CPU- and GPU-models.