Intro

To our knowledge we’re the only company in the world that can perform face recognition, 3D liveness check, deepfake detection, identity concealment check and stream injection (virtual camera) detection from a single image. Our implementation is Passive/Frictionless and takes less than 20 milliseconds.

The stream injection module checks whether the image or video is from a local file or delivered using a virtual camera (OBS Studio, ManyCam, Unity Virtual Camera…). It works on all platforms (Android, iOS, Windows, Linux…). It also detects when a real stream (from real camera) is altered to modify the face (e.g. deepfake).

Deepfake check module detects whether the face is generated ex-nihilo (synthetic, AI-generated) or injected via virtual camera and superimposed.

Identity concealment detects when a user tries to partially hide his/her face (e.g. 3D realistic mask, dark glasses…) or alter the facial features (e.g. heavy makeup, fake nose, fake beard…) to impersonate another user.

The liveness module supports multiple form of spoofing attacks: Paper Print, Screen, Video Replay, 3D (silicone, paper, tissue…) realistic face mask, 2D paper mask…

A facial recognition system without a robust liveness detector is a playground for fraudsters.

Next video shows a stress test on our implementation using different type of 2D/3D attacks:

Contents: