Idemia Claims Fairest Face Verification Among Most Accurate Algorithms in Latest NIST Test

Idemia Scores in Latest Vendor Facial Recognition Test for individual verification show nearly identical false match rates across different demographics, surpassing the 100 most accurate algorithms in terms of fairness, according to an announcement from the company.

The Idemia-009 algorithm scored fourth best in Mugshot, Border Photos and Kiosk Photos categories, seventh best in ‘VisaBorder Photos’ and ninth best in Mugshot Photos after 12 years in the September 1 update to the National Institute for Standards and Technology.

The company, however, seems most proud of its results in terms of demographic differentials, or biases. Idemia claims that its algorithm was found to be more than twice as accurate as the 20 most accurate competing algorithms, according to the Aug. 30 report.

“NIST’s FRVT results are further evidence of the highest standard we have set with our suite of facial recognition technologies, positioning fairness as a key criterion, in addition to accuracy,” says Jean-Christophe Fondeur. , CTO of Idemia. “By being more than double the equity of the most accurate 20, we continue to lead the industry in terms of social responsibility. Idemia is leading the way in the “battle for fairness” and I would like to congratulate our teams of experts for their excellent work in this priority. »

The top-performing algorithms evaluated by NIST in the biometric verification accuracy test remain largely unchanged from recent reports.

Idemia also boasts top marks for accuracy in NIST’s one-to-many test, a top score for single-eye accuracy in the IREX 10 iris biometric benchmark, and a high score in several categories of the test. Fingerprint Recognition PFT III.

Article topics

precision | algorithms | biometric tests | biometrics | demographic equity | Facial Recognition Testing (FRVT) | facial recognition | IDEMIA | NIST

Sharon D. Cole