US 11,670,069 B2
System and method for face spoofing attack detection
Denis Timoshenko, New York, NY (US); Konstantin Simonchik, New York, NY (US); and Alexey Khitrov, New York, NY (US)
Assigned to ID R&D, Inc., New York, NY (US)
Filed by ID R&D, Inc., New York, NY (US)
Filed on Feb. 6, 2020, as Appl. No. 16/783,465.
Prior Publication US 2021/0248401 A1, Aug. 12, 2021
Int. Cl. G06K 9/00 (2022.01); G06V 40/40 (2022.01); G06V 40/16 (2022.01)
CPC G06V 40/40 (2022.01) [G06V 40/172 (2022.01)] 10 Claims
OG exemplary drawing
 
1. A method for face spoofing attack detection, the method comprising:
capturing an image and determining presence of a face in the captured image;
determining a context of the captured image comprising the face, wherein the context relates to a spoofing attack;
highlighting a part of the captured image to generate an image with the highlighted part, wherein the highlighted part is a potential spoofing region determined based on the context of the captured image;
combining the captured image with the image with the highlighted part to generate a summarized image, wherein determining the context of the captured image, highlighting the part and combining the captured image with the image with the highlighted part are performed by a spoofing mask module;
performing face spoofing attack detection in the summarized image, wherein only the highlighted part has priority in the face spoofing attack detection, compared with the rest of the summarized image;
generating a general spoofing score as a result of the face spoofing attack detection based on the highlighted part and the rest of the summarized image, wherein performing face spoofing attack detection and generating a general spoofing score are performed by a liveness module trained separately from the spoofing mask module, and the liveness module and spoofing mask module are independent feature extraction modules;
comparing the general spoofing score result with a threshold for face liveness detection.