The 2022 Biometric Technology Rally, conducted by the Department of Homeland Security, found that the group size of people being captured by face matching software through a border control-like environment had no detrimental impact on accuracy.
The undertaking of this type of study is important to determine whether busy environments like airports with a dense and continuous flow of passengers passing through security touchpoints can have an impact on the accuracy of identity verification with singular face matching.
Testing of many vendor systems took place last autumn which disclosed that our of 40 different configurations of 10 face matching algorithms and four cameras, 17 achieve the top ranking for identification while 31 combinations achieved the set goal of more than 99% accuracy.
575 participants took part in the biometrics rally which represented a wide cross-section of demographic groups and both men and women to whether the face recognition capabilities mitigated bias.
29 variations of the algorithms also successfully avoided capturing live images of people who opted out of testing and bystanders, meeting the privacy goal.
97% of people who were processed in small groups took less than two seconds to be recognised accurately.
The only technical fault of the cameras was to supply/capture a photograph of the individuals which algorithm matching.
Arun Vemury, Center Lead, DHS S&T Biometric and Identity Technology Center said the purpose of the rally was to raise the “performance bar, to further challenge industry to develop technologies that address operational challenges and constraints” such as accuracy, speed, and ease of use while also “addressing privacy and civil rights and civil liberties concerns”.
