Thursday, March 19, 2020

Polarization Imaging for FaceID Anti-Spoofing

Arxiv.org paper "Face Anti-Spoofing by Learning Polarization Cues in a Real-World Scenario" by Yu Tian, Kunbo Zhang, Leyuan Wang, and Zhenan Sun from Institute of Automation, Chinese Academy of Sciences, proposes polarization sensor for fake face detection:

"In this paper, we propose a polarization face anti-spoofing (PAAS) method using a deep learning computational model to learn physical cues in polarization images. Because polarization reveals the information of shape, material, roughness and other attributes of an object, it is difficult to imitate or change. Combining optimized CNN algorithm with the new on-chip four-directional polarization image sensor, our PAAS system is able to realize realtime detection of a variety of face spoofing attacks in a real-world scenario."

No comments:

Post a Comment

All comments are moderated to avoid spam and personal attacks.