YooniK places among facial recognition accuracy leaders from Western nations in December NIST test
Face biometrics startup YooniK scored accuracy results in the mid-December edition of NIST’s Face Recognition Vendor Test (FRVT) that the company says put it among the leading biometrics vendors from Western countries.
The company was founded in 2019, and placed in the top 50 among 260 entrants in both visa categories and both mugshot categories in the last FRVT of 2020. On the updated leaderboard, the company’s ‘yoonik-001’ ranks 33rd in visa photos and 28th and 29th in accuracy matching against the mugshot datasets.
YooniK provides the cloud-based YooniK Face solution as a subscription service for a wide range of consumer-facing industries. The company’s mission is to simplify consumer interactions both on an offline with contactless, secure and private face biometric authentication to any IoT device, a company representative told Biometric Update in an email. The company plans to bring “unprecedented privacy levels to the market” for companies in retail, banking and payments, travel and hospitality, healthcare, self-service and delivery.
With a 33rd-place finish in the two categories for mugshot photos based on false non-match rate (FNMR) at a false-match rate (FMR) of 0.00001 and for mugshots after 12 years, YooniK ranked behind only Idemia and Paravision among North America- and Europe-based entrants.
“We’re on a mission to give time and freedom to people. We believe YooniK will offer one year of free time across a person’s lifetime, which is currently wasted in inefficient bureaucracy for someone to check who you are or that you can access some service: to make a payment, open an account, onboard to some platform, check into a hotel, buy prescription medicines, or instantaneously prove loyalty on-the-go,” says YooniK Co-founder and CEO Pedro Torres. “In a world where time is our most valuable resource, just imagine what you could do with that extra year.”
This post was updated at 9:12am Eastern on January 22, 2021 to include the latest FRVT results for the ‘yoonik-001’ algorithm.