AU10TIX Releases Deep-Learning Based ID-to-Selfie Face Comparison Beta

Advanced Deep Learning algorithm enables AU10TIX face comparison service to compare quicker and more accurately face photos retrieved from ID documents with selfie pictures captured by device cameras. AU10TIX Deep Learning technology helps overcome the problematic of conventional Face Comparison in handling image variations and image quality issues.

AU10TIX R&D team implemented state-of-the-art Deep Learning algorithmics in biometric identification. The company’s Document-to-Selfie face comparison algorithm has proven exceptionally well accuracy and speed in comparing face images captures from ID document image files to selfie pictures taken by device cameras such as smartphones. AU10TIX has extensive experience in face comparison algorythmics since its early days as an airport border control security technology company. But when exploring solutions for online and mobile Customer-Not-Present situations the company’s R&D team established that conventional algorithms are neither quick nor effective enough in handling face comparison with pictures taken in rudimentary conditions, available in different qualities and captured in different times. AU10TIX’s new Deep Learning based face comparison beta has already been tested extensively, showing so far clear superiority over existing solutions.

“Identity is not just about ID documents any more”, says Ron Atzmon, Managing Director of AU10TIX: “Biometrics is becoming standard and AU10TIX is there to make both work better than before. Face comparison is not a new thing. It’s here for a long time now. But what works when a customer is in front of you doesn’t necessarily work in a Customer-Not-Present situation. Conventional technology has its limits. Everyone who tested it knows the facts. The smart things about the new algorithm is its ability to compare a live selfie with an ID document picture. These two pictures have been taken in different times and different conditions and are obtained from different sources. Service providers can now gear up for stronger customer screening yet streamlined onboarding process.”

“The life of fraudsters just got harder”, says Ofer Friedman, VP Marketing of AU10TIX: “In a market where the vast majority of fraud cases are based on genuine stolen data, document authentication and biometrics become simply a must! Now, fraudster have to work even harder: Not only do they have to pass the only ID authentication platform with multi-factor, forensic level authentication, but now they must pass face comparison as well… This is just a tip of the iceberg of what’s coming.”