Unissey claims 18 percent face biometrics performance gain in latest NIST testing

Unissey has achieved an 18 percent improvement in the performance of its face biometrics algorithms in the global benchmark test, according to a company announcement.

The company first submitted its biometric algorithms to the United States’ National Institute of Standards and Technology (NIST) for evaluation in the Face Recognition Vendors Test (FRVT) in July, 2021, and the company says it achieved an overall result in the top half of the field.

Unissey’s algorithms entered on November 29, 2021 scored a false non-match rate (FNMR) of 1.47 percent at a false match rate (FMR) of 1 in a million in comparisons against the ‘visa-border’ dataset in the latest FRVT 1:1 Verification report, published December 16.

The company contrasts this performance with that with human examiners, who are found to accept false matches for up to 30 percent of facial image comparisons.

Unissey recently partnered with Édokial to provide selfie biometric authentication and passive liveness detection for the latter’s KYC offerings.

Article Topics

accuracy  |  algorithms  |  biometrics  |  face biometrics  |  identity verification  |  NIST  |  testing  |  Unissey