.. _new:

What's New in FindFace Multi 2.2
======================================


.. rubric:: New Features:

* :ref:`Maps <enable_map>`: see geographic location of cameras with their statuses on the map.
* :ref:`alarm_feed`: an alarm management tool that displays alarms once they occur; allows adding a comment and downloading a report. 
* :ref:`Alert manager <configure_alert>`: a simple tool to configure an alert and start receiving alarm notifications into selected delivery channels.
* Support of :ref:`Telegram bot <configure_tg_bot>` as a delivery channel for receiving alarm notifications.
* :ref:`Batch upload of vehicle records from a CSV file <batch-upload>`.

.. rubric:: Technical Changes:

* FindFace core update: FF.Server 11.240325 version inside FindFace Multi 2.2.


.. rubric:: Enhanced Algorithms, UI, UX:

* New :ref:`License <licensing-principles>` interface: work with multiple licenses at once, combine license limits.


.. rubric:: Fresh Neural Networks:

+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Type                  | Model                               | Description                                                                    |
+=======================+=====================================+================================================================================+
| Face detection        | ``facedet.kali.005``                | a new face detection model                                                     |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Body detection        | ``bodydet.gustav_normal.021``,      | new body detection models that have been trained on more complex cases if      |
|                       | ``bodydet.kali.021``,               | compared to a predecessor                                                      |
|                       | ``bodydet.kali_accurate.021``       |                                                                                |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Multiclass face/body  | ``bodydet_facedet.kali.001``,       | new multiclass detection models that allow using one detector for face and     |
| detection             | ``bodydet_facedet.kali_normal.001`` | body instead of two separate ones, which saves resource usage                  |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+         
| Vehicle detection     | ``cardet.kali_accurate.008``        | a new vehicle detection model                                                  |
|                       +-------------------------------------+--------------------------------------------------------------------------------+
|                       | ``car.gustav_normal.007``           | an updated version of the vehicle detection model                              |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Face recognition      | ``nectarine_l_320``                 | the default face biometric model upon a clean installation, more accurate and  |
|                       |                                     | slightly faster than previously used ``mango_320`` model                       |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Face attribute        | ``faceattr.age.v3``,                | updated versions of face attribute recognition models with improved            |
| recognition           | ``faceattr.gender.v3``,             | characteristics                                                                |
|                       | ``faceattr.liveness_pacs.v3``,      |                                                                                |
|                       | ``faceattr.liveness_web.v1``,       |                                                                                |
|                       | ``faceattr.quality.v5``             |                                                                                |
|                       +-------------------------------------+--------------------------------------------------------------------------------+
|                       | ``faceattr.glasses4.v0``            | a new model that predicts whether a person wears glasses; four classes are     |
|                       |                                     | supported: medical glasses, sunglasses, other (e.g., carnival glasses, sports  |
|                       |                                     | glasses, etc.), none (no glasses found on a person’s face)                     |
|                       +-------------------------------------+--------------------------------------------------------------------------------+  
|                       | ``faceattr.eyes_attrs.v0``          | a new model that recognizes whether person's eyes are opened/closed or not seen|
|                       +-------------------------------------+--------------------------------------------------------------------------------+
|                       | ``faceattr.liveness_mobile.hart``   | a new liveness recognition model for mobile phones that is an improved version | 
|                       |                                     | of ``liveness.goodwin`` model (``liveness.goodwin`` passed iBeta Level 2 test) |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Vehicle image         | ``carnorm.briacon.v2``              | an updated vehicle image normalization model that is trained on more data      |
| normalization         |                                     | if compared to previously used ``briacon.v0`` model                            |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Vehicle attribute     | ``carattr.categories.v1``,          | updated versions of vehicle attribute recognition models with improved         |
| recognition           | ``carattr.description.v1``,         | characteristics                                                                |
|                       | ``carattr.quality.v1``,             |                                                                                |
|                       | ``carattr.license_plate_quality.v3``|                                                                                |
|                       +-------------------------------------+--------------------------------------------------------------------------------+
|                       | ``carattr.license_plate.v9``        | an updated version of the license plate recognition model                      |
|                       |                                     | that supports more countries (including China, Egypt, Japan,                   | 
|                       |                                     | Mongolia, Morocco, Palestine, Singapore, South Korea, Bahrain, Colombia,       | 
|                       |                                     | Peru, Qatar, Tunisia, Türkiye) and regions of Thailand                         |
|                       +-------------------------------------+--------------------------------------------------------------------------------+
|                       | ``carattr.special_types12.v0``      | a new model that includes recognition of one more special vehicle type:        |
|                       |                                     | cash-in-transit vehicles                                                       |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+
| Body attribute        | ``pedattr.fall.v1``                 | a new model that predicts whether a person is falling (has fallen) or not      |
| recognition           +-------------------------------------+--------------------------------------------------------------------------------+
|                       | ``pedattr.handface.v1``             | a new model that predicts whether a person is smoking, using a phone, or       |
|                       |                                     | calling                                                                        |
+-----------------------+-------------------------------------+--------------------------------------------------------------------------------+