IBIA: New NIST report on face recognition a 'game changer'
14 February 2020 17:58 GMT

The International Biometrics + Identity Association has said a recent NIST report on the performance of facial recognition algorithms across different demographics is a "game-changer".

IBIA has said in a paper summarising NIST's findings that the latter's new report provides vital new data on the performance of algorithms across demographic groups.

"This data serves to debunk the semantically-loaded misleading arguments on facial recognition performance that privacy activists have pushed in their zeal to ban a technology that enhances public safety and security. "

The NIST testing showed wide variations in performance, ranging from algorithms that are less accurate than a coin toss, to highperforming algorithms that are overwhelmingly accurate with virtually undetectable demographic differences. 

These latter algorithms are 20 times more accurate than the most highly-skilled human groups.

The IBIA notes that unfortunately, media coverage drew generalizations from a few algorithms and focused on the low-performing algorithms in the study, not the range of algorithms tested, and certainly not the significant number of high-performing algorithms.

"This paper summarizes the key NIST findings of high-performing algorithms (ignoring those algorithms that no prudent organization would use); how the data serves to debunk the fundamental misrepresentations of the activist arguments that the algorithms are biased and not good enough."