Cognitec takes leaderboard positions in FRVT tests
12 December 2018 15:10 GMT

German face recognition firm Cognitec has reported remarkable leaderboard positions in the NIST Ongoing Face Recognition Vendor Test (FRVT) 1:1.

It said the November 2018 results, comparing 39 algorithms, show Cognitec’s high performance for accuracy tests with VISA photos. In addition, its algorithm performed best for template generation speed at a false non-match rate below 0.01.

"In general, NIST reports massive gains in accuracy in the last five years, and these far exceed improvements made in the prior period between 2010 and 2013. At least 28 developers’ algorithms now outperform the most accurate algorithm from late 2013. Face recognition has undergone an industrial revolution, mainly due to deep neural network methods, the use of larger databases, and increasing computing power".

NIST advises users to thoroughly evaluate face recognition technologies, noting that leaderboard algorithms are significantly more accurate than most others available on the market.