Startup Unissey passes biometric presentation attack detection test, first by French lab CLR
The Cabinet Louis Reynaud Labs (CLR) group has conducted its first evaluation of the ISO/IEC 30107 standard, finding French biometrics firm Unissey compliant with Level 1 for face biometric presentation attack detection (PAD), ensuring that a person is who he or she claims to be and is alive and present in front of the camera for verification.
CLR Labs recently announced a partnership with the assessment body LSTI to assess remote identity verification providers in terms of biometrics and liveness detection.
The lab confirmed that Unissey’s liveness detection returned no false negative matches after several thousand presentation attacks with paper photo prints, video screens and 3D masks, presenting the biometric data of real, demographically diverse subjects.
“We are very proud to work with CLR Labs and to benefit from their extensive biometric expertise, which is in line with our approach to continuously challenge our liveness detection solution,” said Olivier Roblin, CTO of Unissey.
The firm’s solution is now proceeding to Level 2 evaluation, involving more masks and attacks targeted specifically to attack Unissey’s liveness detection.
A French lab assessing compliance to the ISO standard for the first time, and for a French biometrics startup, is a proud moment for those involved who see it as the emergence of a European ecosystem and strategic autonomy for France and Europe.
Stéfane Mouille, director of CLR Labs said, “Building biometric evaluation and certification capabilities in Europe is a matter of European sovereignty, we are very honored that the company Unissey has decided to trust us to evaluate its solution for liveness detection based on AI algorithms.”
Unissey Computer Vision Engineer Idriss Mghabbar will deliver a presentation on bias mitigation in anti-spoofing in the next EAB virtual lunch talk.
Related Posts
Article Topics
biometric liveness detection | biometrics | CLR Labs | France | ISO standards | presentation attack detection | testing | Unissey
Comments