Cognitec Highlights ‘Optimal Balance’ of Its Facial Recognition Tech After NIST Testing

Cognitec is celebrating a strong performance in the latest round of facial recognition testing by the National Institute of Standards and Technology (NIST).

Cognitec says that its overall results in the NIST Face Recognition Vendor Test illustrate “the optimal balance between speed and accuracy of face matching processes,” as the company put it in a statement. Elaborating on the point, the company said that its solution was “in the best position of all algorithms” when comparing the speed of its biometric template generation to its False Rejection Rate with respect to matching faces against databases of mug shots.

Commenting further, Cognitec Director of Algorithm Develpment Dr. Thorsten Thies highlighted the consistency of the performances delivered by the company’s algorithm across different tests.

“For the test with 12 Million mugshot images, Cognitec achieved rank 24 of 165 algorithms with a match rate of 98.5%,” he said. “In the test with 1.6 Million mugshots, with a 99.4% match rate, we ranked 26 of 299 algorithms. These results show remarkable performance consistency, regardless of database size.”

Cognitec says that the facial recognition algorithm it submitted for this round of testing was a new one featuring an upgraded face detection feature. According to Dr. Thies, the algorithm demonstrated “significant accuracy advances” in comparison to the algorithm submitted for testing earlier this year.

It’s the latest illustration of Cognitec’s ongoing efforts to improve its facial recognition technology. Roughly a year ago, for example, the company announced that it had upgraded its FaceVACS engine to enable it to better recognize individuals wearing face masks – an innovation spurred by the COVID-19 pandemic.

The development also helps to illustrate the importance of NIST’s third party testing programs, which offer a level playing field on which many vendors compete. NIST’s testing extends beyond facial recognition, as well.

–

October 19, 2021 – by Alex Perala