NIST and DHS Biometric Technology Center expecting significant accuracy gains

Officials from the National Institute of Science and Technology (NIST) expect error rates in matching faces to mugshots to be reduced by a factor of three in the new Facial Recognition Vendor Test which it is currently working on, FedScoop reports.

FRVT results already improved by 20 times from 2013 to 2018, which NIST attributed to the development of convolutional neural networks. NIST Computer Scientist Mei Lee Ngan told the audience at a recent Security Industry Association that leading algorithms like Microsoft’s can now match side photos to mugshots with similar accuracy as algorithms in 2010 could match frontal photos to mugshots. There a spread amongst the 140 algorithms that will be tested, however.

“There’s a percentage of [algorithms] that are really good and very capable, but there is this long tail of facial recognition algorithm developers that just don’t do that good of a job,” Ngan said, according to FedScoop. “So the technology hasn’t been commoditized yet.”

The Department of Homeland Security’s Biometric Technology Center (BTC) is also expected to publish results from its second Biometric Technology Rally in the next several weeks. That test put hundreds of volunteers through 12 high-throughput gates provided by participants over a twelve week period earlier this year.

BTC Director Arun Vemury says that some systems failed to capture enough data to deliver readings as often as 20 percent of the time in the first rally. He also notes that the fastest systems were not the most accurate, FedScoop reports.

Iris scans of subjects in motion continue to challenge companies, but the results for the latest test were still promising, according to Vemury.

“Timing got better, failure to acquire rates generally went down … and we saw very good satisfaction scores.”

Customs and Border Protection (CBP) Acting Director of Biometric Exit Strategic Transformation Jeni Best says that the agency plans to integrate its Traveler Verification Service with the Transportation Security Administration’s (TSA’s) Secure Flight program. She also says fingerprint and iris biometric checks are not feasible in environments in which operators must work quickly, saying, “we can’t impede the traveling public.”

A new Reservations.com survey suggests that only 43 percent of Americans approve of facial recognition use in airports, with nearly one in three against it. Biometric Update recently discussed the state of airport biometrics and efforts to align them with several industry experts.

Article Topics

accuracy  |  biometric testing  |  biometrics  |  DHS  |  facial recognition  |  NIST