Video-based facial recognition - Gemalto Cogent Live Face Identification System


Last updated 31 August 2018

Video-based facial recognition system

Gemalto Cogent Live Face Identification System is a video-based facial recognition system that automatically recognizes faces in a crowd-- even in dynamic, uncontrolled environments-- and sends real-time alerts so you can take action quickly.​

The system can be integrated into a wide range of video equipment, and advanced algorithms increase the accuracy of matches.

Facial recognition  

Face detection in a crowd in real-time

Gemalto Cogent Live Face Identification System (LFIS) includes 2 major components

  • ​Core LFIS and LFIS Check SDK (Software Development Kit). Core LFIS provides video based face recognition designed for face detection in a crowd in real-time or post-event and searched against a built-in person of interest list.
     
  • LFIS Check SDK is a robust software development kit that allows developers to create applications that use face as a biometric identifier.

    The SDK comes with a demonstration app that shows how the SDK can be integrated with a Gemalto document reader to match live faces with faces from documents. Our live face recognition solution has been designed to be scalable and is built on top of a configuration of stable technologies.

     Traditionally, large scale distributed biometric systems require highly experienced product specialists to configure. LFIS has a convenient configuration system and a rich set of RESTful (representational state transfer) web services.

2018 face recognition benchmarks

DHS Rally

At the end of May 2018, the US Homeland Security Science and Technology Directorate published the results of sponsored tests at the Maryland Test Facility (MdTF) done in March. These real-life tests measured the performance of 12 facial recognition systems in a corridor measuring 2 m by 2.5 m. 

Gemalto, also known as 'Castle' in the anonymized results (organization is gemalto) shared by the sponsors performed exceptionally well.

Gemalto's solution utilizing its facial recognition software (LFIS) achieved excellent results with:

  • a face acquisition rate of 99.44% in less than 5 seconds (against an average of 68%), 
  • a Vendor True Identification Rate of 98% in less than 5 seconds compared with an average 66%, 
  • and an error rate of 1% compared with an average 32%.

​ NIST FRVT 1:1 test

In the most recent National Institute of Standards and Technology (NIST) FRVT 1:1 test (last update June 2018) Gemalto Cogent algorithm was the first non-Chinese/Russian vendor in the mugshot test and VISA test (and respectively #4 and #6 vendor).

The mugshot and VISA test are the test using the largest photo test set in the NIST FRVT test suite (100K and 1M respectively).​

​​​​4 ways it can advance security from reactive to proactive

Proactive Security​ 

Proactive Security​

• Automatically & simultaneously recognizes multiple faces in a crowd
• Greater accuracy of face matches with "several-to-many" comparison
• Take action quickly with near real-time alerts delivered to mobile devices or connected PC's​


 
Face identification  

Easily Scalable

• No proprietary hardware required 
• Images can be captured from a wide range of compatible cameras
• Easily Scalable– can support large scale systems and large database sizes​


 ​​
Efficiency with accuracy​ 

Efficiency with Accuracy​

•  Accurate capture– supports multiple faces per frame, multiple frames per second speed, and multiple resolutions
• More effective– advanced face recognition algorithms make the software more effective in identifying matches from low-quality images


 
Flexible software

Flexible Software

• Dashboard watch list– can enroll & categorize over 1 million faces
• Investigation support– import footage and still images to help identify suspects with the "several-to-many" comparison
• Software Developer's Kit (SDK)– algorithms are available in a robust software development kit to integrate 1:1 matching with Gemalto document readers​​.


 
​​Discover top facial recognition news and trends in our August 2018 web dossier,​

 

 Documents