Cognitec reports biometric identification success in NIST facial recognition testing

The latest results from NIST’s Face Recognition Vendor Test (FRVT) 1:N include successful participation by Cognitec, which according to a company announcement confirm the solid market position of its FaceVACS matching technology, and the achievements of its algorithm research and development.

The latest NIST FRVT 1:N test results includes 203 algorithms submitted by 51 commercial developers and a university. Cognitec says they show the excellent identification performance of its technology for large mugshot galleries, as well as greatly improved performance in matching unconstrained images produced in video applications which have severe variations in pose.

In tests of mugshot searches for identification, a pair of Cognitec algorithms ranked in the twenties.

Microsoft and NEC topped several categories, while algorithms from Yitu, VisionLabs, and SenseTime were at or near the top in certain tests.

The B12 algorithm submitted to NIST by Cognitec has been integrated into its product portfolio, and the company says it continues to optimize its matching algorithm to balance accuracy and speed in actual products. Cognitec also updated the FaceVACS Engine in August, boosting the performance of its Nero MediaHome platform.

Article Topics

accuracy  |  algorithms  |  biometric testing  |  Cognitec  |  Face Recognition Vendor Test (FRVT)  |  facial recognition  |  NIST  |  research and development