IDEMIA’s Facial Recognition Algorithm Comes Out Ahead in NIST Testing

IDEMIA can boast that its improved facial recognition algorithm is the best available to US government agencies thanks to new testing by the National Institute of Standards and Technology (NIST). The latest NIST Facial Recognition Vendor Test (FRVT) evaluated 127 algorithms from 39 different developers, and found that IDEMIA’s solution outperformed those of similar US government providers by 30 percent in one-to-many facial recognition scenarios.

“IDEMIA has a long history of developing highly accurate, scalable and rapidly deployable biometric technology,” said Donnie Scott, IDEMIA’s Senior Vice President of Public Security for North America. “NIST’s FRVT results are further evidence of the new standard we have set with our suite of facial recognition technologies.”

The FRVT testing specifically measured an algorithm’s ability to compare one face against a large database and return positive matches. IDEMIA’s solution performed best when dealing with standard portraits and Wild Image scenarios, which test images from non-compliant individuals. That means that IDEMIA was able to deliver results even it didn’t have a clear photo of someone who was trying to avoid having their picture taken.

This is not the first time that IDEMIA has bested its competitors in NIST testing. IDEMIA’s facial recognition tech was the most accurate selfie solution in a previous round of FRVT testing in 2018, while the company itself recently formed a partnership with Onfido to enhance its selfie authentication services.

The results come soon after IDEMIA’s announcement that it has teamed up with Infineon and Pmdtechnologies to further refine its 3D facial recognition technology.

–

March 27, 2019 – by Eric Weiss