HyperVerge is making a splash in the NIST’s most recent Face Recognition Vendor Test (FRVT). The facial recognition developer took the seventh overall spot in the organization’s FRVT 1:1 evaluation, and climbed all the way to second in the Border Images Benchmark test. HyperVerge noted that the Border Images Benchmark is one of the more difficult challenges for most developers, since the quality of the images varies and the algorithms are forced to match faces at a wide variety of angles.
The company performed similarly well with regards to 1:N facial matching. HyperVerge held the eighth position on the global 1:N Identification Leaderboard, and climbed all the way to fifth and sixth, respectively, in the 1:N Investigation Track and the NIST Mugshot benchmark. 1:N matching solutions are often used for fraud prevention during customer onboarding, and for deduplication when used with passport databases and other large-scale registries.
“Accuracy and reliability of face recognition are very important to our clients, and we at HyperVerge are constantly innovating to build better algorithms,” said HyperVerge Co-Founder and CTO Vignesh Krishnakumar. “Recent rankings on the NIST leaderboard prove the robustness of our systems.”
HyperVerge went on to argue that its solution is well-suited to identity verification and fraud prevention use cases. The company suggested that automated identity tech can help organizations reduce their operational costs, insofar it mitigates the need for manual identity verification processes. It also enables a faster onboarding process for individual end users.
Experts have noted that demand for remote onboarding solutions has skyrocketed in the past few years, primarily as a result of the shift to remote and hybrid work environments during the pandemic. The broader facial recognition market is expected to exceed $10 billion within four years, and Emergen Research predicts that it could climb as high as $13.87 billion leading up to 2028.
May 30, 2022 – by Eric Weiss