Japan’s tech giants, Facephi, Innovatrics climb NIST face biometrics ranks

The mix of companies represented on the leaderboard of the U.S. National Institute of Standards and Technology’s Face Recognition Technology Evaluation has shifted again with the latest update to its biometric verification report and leaderboard. Japanese tech giants have taken prominent spots on the first page of the leaderboard, Facephi sees a competitive advantage in its tolerance of challenging conditions and Innovatrics has made another strong showing.

NIST updated its FRTE 1:1 Verification track on March 18.

Facephi hailed its strong results for accuracy and efficiency in a release after placing sixth (on the “one per developer” list) in the category for Border to Visa checks at a yaw of up to 45 degrees (ie. with turned faces) out of 390 algorithms, with a false non-match rate (FNMR) of 0.0043. The company’s algorithm submitted on February 11 also ranked in the top 15 comparing mugshots over a timespan of up to 12 years, and the top 30 in the straight mugshot comparison category.

Facephi says its results show the exceptional performance of its face biometrics compared to competitors in challenging real-world scenarios, like age variation detection and facial recognition at angles up to 45 degrees.

“These achievements reaffirm Facephi’s commitment to continuous innovation and excellence in digital protection,” says CEO Javier Mira. “We strive to provide our clients with faster, more accurate identity verification solutions that reduce fraud risk and enhance the user experience. These outstanding results further solidify our position as a market leader, offering biometric solutions that adhere to the highest international standards.”

Other prominent entries since ROC submitted an algorithm on January 31 that stood out amongst Western developers, the day after the first FRTE 1:1 update of the year, include Innovatrics, Samsung, Toshiba, VinBigData and Deepsense.

Innovatrics came second and fifth in the mugshot categories noted above, respectively, and in the top 10 across the board (excluding Visa-Visa comparisons, for which it was not assessed).

Toshiba finished in the top ten in three categories, and Samsung’s algorithm made the top 30 across all six categories it was assessed in.

VinBigData was 21st or higher across the same six categories, and Deepsense made the top 15 in four categories.

Article Topics

biometric testing  |  biometrics  |  Face Recognition Technology Evaluation (FRTE)  |  FacePhi  |  facial recognition  |  Innovatrics  |  NIST