Documentation for Face Intellect 8.0-8.1. Documentation for other versions of Intellect is available too.

Previous page Configuring the program modules  Face detection Next page

Note

Face recognition modules can operate in multi-threading mode. By default, all available processor cores are used. The number of processor cores actually used in the operation of the face recognition module is set using the CpuCoreCount registry key (see Registry keys reference guide).

To enable the face recognition, it is necessary to activate one of the software modules:

  1. Cognitec recognition module.
  2. VisionLabs recognition module. On the basis of the created object, it is also necessary to create an object representing the number of reference faces in the database.
  3. Tevian recognition module. On the basis of the created object, it is also necessary to create an object representing the number of reference faces in the database.
  4. STC recognition module.
  5. Face recognition VideoIntellect 1.1.

To activate the software module, it is necessary to create an object on the basis of the Face Recognition server object with the same name as the required module.

The following objects can also be created on the basis of the Face Recognition Server object:

  1. The Archive search channel module which is used for searching the captured face in the face database and viewing the video archive of the captured face.

  2. The Face characteristics recognition channel module which is used for the face characteristics recognition for a specified channel.

    Attention!

    The Face characteristics recognition channel module can only operate with the Tevian or VideoIntellect 1.1 face recognition module.

    Note

    If the Face characteristics recognition channel is not activated, then the face characteristics of all newly captured faces will be marked as Undefined in the Face recognition and search interface window. However, if this module is activated later, then as Face Intellect operates and accumulates data, the undefined face characteristics of already captured persons will be changed to averaged values.

    This module provides the recognition of the following face characteristics:

    • age;

      Note

      If the age of the same person is determined with poor accuracy, then it is necessary to set the 1 or 5 value for the Tevian.TrackBestFramesCount key (for details, see Registry keys reference guide, for information about working with the registry, see Working with Windows OS registry).

    • gender;

    • emotion;

    • glasses;

    • facial hair;

    • hair color;

    • type of bald head;

    • headwear;

    • artificial face;

    • face concealment.

      Note

      If masked faces are not recognized, then it is necessary to set the 0 value for the Tevian.GlobalTrackingDetectionFilter registry key (for details, see Registry keys reference guide, for information about working with the registry, see Working with Windows OS registry).

  3. The Visitor counting channel module which is used for counting visitors and repeated passes. 

  4. The Face database replication module which is used for the Server face database replication in a distributed configuration (see Face database replication).
  5. The Recognition channel and Tevian/Cognitec/VisionLabs/STC/VideoIntellect 1.1 recognition modules which are used for real-time face recognition for the specified channel.
  6. The Captured faces processing channels module is used for capturing faces and determining their temperature in real time for the selected video channel. To use this channel, it is necessary to configure the Face detection to work with a thermal camera (see Configuring the Face detection module combined work with thermal camera or external system).

Attention!

After the Visitor counting channel, Archive search channel, Recognition channel, Face characteristics recognition channel and Captured faces processing channels modules are activated, it is necessary to select the required Face detection object on the settings panel of the corresponding object (see Configuring the captured faces recognition parameters).

  • No labels