QazSmartVision shows out in NIST face biometrics, image quality testing

Biometrics providers from in and around the Middle East are taking a more prominent place in facial recognition evaluations by the National Institute of Standards and Technology. Developers from Kazakhstan and the UAE are among the accuracy leaders, while a Turkish developer has joined the program for the October 7 update of the ongoing Face Recognition Technology Evaluation (FRTE) 1:1 Verification.

Kazakhstani startup QazSmartVision.AI submitted an algorithm in September which scored the highest accuracy in the mugshot-mugshot, visa-border, border-border and border-kiosk categories. The result follows a strong showing with an algorithm submitted earlier this year.

Other developers retaining spots among the most-accurate algorithms include UAE-based Recognito, Cloudwalk, Viante, Megvii, SenseTime and Idemia.

New entrants to the 1:1 track are Turkish developer Enqura, U.S.-based Kloudspot, which has an office in Dubai, and Polish developer Pentacomp Systemy Informatyczne.

An October 2 update of the Face Analysis Technology Evaluation (FATE) Quality track also shows strong performance by QazSmartVision’s quality assessment algorithm. The best false non-match rate after removal of the 5 percent of images with the lowest quality was by a secunet algorithm, while one from Intema showed the best FNMR result after removing the lowest quality 1 percent of images.

Idemia’s latest submission, along with QazSmartVision’s, were in the top 5 by FNMR for both categories.

Article Topics

accuracy  |  biometric testing  |  biometrics  |  Face Recognition Technology Evaluation (FRTE)  |  facial recognition  |  NIST  |  QazSmartVision.AI