Alice gets high scores for PAD, face biometric ID verification
Alice Biometrics, which provides identity verification based on face biometrics, is among a handful of providers to post positive results in the U.S. National Institute of Standards and Technology (NIST) assessment of 82 passive facial presentation attack detection (PAD) algorithms. The test is NIST’s first-ever evaluation conducted on PAD.
A release from Alice, which is based in Pontevedra, Spain, says the results further position the company as a global leader in passive facial liveness detection and remote identity verification.
“NIST is globally recognized for its expertise in evaluating commercial biometric algorithms, and we greatly appreciate their diligent evaluation of PAD,” says Esteban Vázquez, the CTO and co-founder of Alice Biometrics. “We are thrilled with the results, as they validate our belief that biometric security should be seamless. These results provide compelling evidence that the current generation of liveness solutions can offer both robust security and user convenience.”
The NIST evaluation, which evaluates software-based presentation attack detection from 2D images and videos, assessed algorithms from 45 developers providing impersonation and evasion PAD. Testers deployed silicon masks, photo print and replay impersonation attacks, which together account for 98 percent of attack attempts, according to the announcement. Alice participated only in testing for impersonation. Scores are based on measuring error rates for both convenience (in terms of false rejections) and security (false acceptances) points.
For realistic silicone mask presentation attack testing, Alice was one of only three algorithms to reject fewer than five percent of genuine users (BPCER > 0.05), a threshold over which use in production environments is not viable.
NIST Internal Report 8491, Face Analysis Technology Evaluation Part 10: Performance of Passive, Software-Based Presentation Attack Detection (PAD) Algorithms is available in full here. Alice’s “distillation” is available here.
Alice’s PAD kudos follow NIST remote ID verification ‘win’ in July
Alice is enjoying a run of success with NIST testing. In July, results from the NIST’s one-to-one test for facial recognition providers put the company among the leading Western developers for remote identity verification using face biometrics.
According to a blog post on Alice’s website, that evaluation used the VISABORDER test scenario, in which identity is verified through the comparison of two biometric factors with a set of face photographs specially selected to measure the performance of remote video identification (VDI) systems. The company claims that it scored the best result in this category among remote video-identification developers from around the world.
Alice posted a False Non-Match Rate (FNMR) of 0.3 percent and a False Match Rate (FMR) of one in a million verifications.