Face Identification

Identification: chooses from a list of biometric models which one corresponds to an image

  • Face identification able to work in closed set and open set mode
    • closed set: claimed person is in the database
    • open set mode: claimed person may not be in the database
  • Face identification algorithm can be chosen according to your needs:
    • real time: standard algorithm for real time calculation on a video stream with a standard PC
    • mobile: optimized algorithm for embedded devices; less accurate but quicker
    • precise: slower but more precise algorithm


Security levels can be chosen according to your security needs.

Face identification can be used in a lot of different use cases. In car industry for example, it allows to identify the driver, adjust the seat position and apply specific settings (audio, steering wheel,...) or even collect statistics.

Computing Time and Memory Information

  Computer ("real time" algorithm) Mobile ("mobile" algorithm)
Face Localization 17 [ms] -> 58 fps 17 [ms] -> 58 fps
Face Verification 27 [ms] -> 37 fps 30-90 [ms] -> 11-33 fps
Face Identification (with 20 identities to check) 62 [ms] -> 16 fps  


The library is robust to low resolution, slight light variation and face occlusion. It is also robust to beard variations, glasses changing, skin tone change ... Heavy side face lighting can compromise recognition. The background has no influence for face recognition.

The library can be embedded in multi-threaded applications too. This SDK have been used in several applications for years and all known bugs have been corrected.

Sample Codes

Sample codes for detection and verification are provided along with the library with a clear documentation. Error codes are handled for debugging purpose.

Application Examples

  • Interactive Kiosks
  • Photo Tagging
  • Access Control
  • Smart TV's