US 11,869,255 B2
Anti-counterfeiting face detection method, device and multi-lens camera
Xing Su, Hangzhou (CN); Jiajun Shen, Hangzhou (CN); Hui Mao, Hangzhou (CN); and Shiliang Pu, Hangzhou (CN)
Assigned to HANGZHOU HIKVISION DIGITAL TECHNOLOGY CO., LTD., Hangzhou (CN)
Appl. No. 17/283,920
Filed by Hangzhou Hikvision Digital Technology Co., Ltd., Hangzhou (CN)
PCT Filed Oct. 11, 2019, PCT No. PCT/CN2019/110652
§ 371(c)(1), (2) Date Apr. 8, 2021,
PCT Pub. No. WO2020/073993, PCT Pub. Date Apr. 16, 2020.
Claims priority of application No. 201811192155.4 (CN), filed on Oct. 12, 2018.
Prior Publication US 2021/0397817 A1, Dec. 23, 2021
Int. Cl. G06V 20/64 (2022.01); G06T 7/55 (2017.01); G06V 40/16 (2022.01); G06V 40/40 (2022.01); H04N 23/11 (2023.01); G06V 10/60 (2022.01); G06V 10/143 (2022.01); G06V 10/46 (2022.01); G06V 10/58 (2022.01)
CPC G06V 20/64 (2022.01) [G06T 7/55 (2017.01); G06V 10/143 (2022.01); G06V 10/60 (2022.01); G06V 40/166 (2022.01); G06V 40/45 (2022.01); H04N 23/11 (2023.01); G06T 2207/10048 (2013.01); G06T 2207/20081 (2013.01); G06V 10/467 (2022.01); G06V 10/58 (2022.01)] 8 Claims
OG exemplary drawing
 
1. An anti-counterfeiting face detection method, which is applied to a multi-lens camera comprising a Time-Of-Flight (TOF) camera and a Red-Green-Blue (RGB) camera, and comprises:
acquiring a depth image, an infrared image and an RGB image by using the TOF camera and the RGB camera;
analyzing the RGB image through a preset face detection algorithm to determine an RGB face region of a face in the RGB image and position information of the RGB face region;
determining a depth face region of the face in the depth image and an infrared face region of the face in the infrared image based on the position information of the RGB face region; and
determining that the face passes the detection when the depth face region, the infrared face region and the RGB face region meet corresponding preset rules respectively;
wherein determining that the face passes the detection when the depth face region, the infrared face region and the RGB face region meet corresponding preset rules respectively, comprises:
calculating a similarity between the depth face region and a reserved priori face template;
determining a first image feature of the infrared face region, and analyzing the first image feature through a pre-trained first deep learning model to obtain an infrared analysis result;
determining a second image feature of the RGB face region, and analyzing the second image feature through a pre-trained second deep learning model to obtain an RGB analysis result; and
determining that the face passes the detection, when the similarity is greater than a preset similarity threshold, the infrared analysis result indicates that the face is a living body, and the RGB analysis result indicates that the face is a living body;
wherein calculating the similarity between the depth face region and the reserved priori face template comprises:
determining a side view and a top view of the face based on the depth face region; and
comparing a face depth change curve of the side view and a face depth change curve of the top view with a face depth change curve of the reserved priori face template to determine the similarity between the depth face region and the reserved priori face template.