TY - GEN
T1 - A 3D Mask Face Anti-Spoofing Database with Real World Variations
AU - Liu, Siqi
AU - Yang, Baoyao
AU - Yuen, Pong C.
AU - Zhao, Guoying
PY - 2016/12/16
Y1 - 2016/12/16
N2 - 3D Mask face spoofing attack becomes new challenge and attracts more research interests in recent years. However, due to the deficiency number and limited variations of database, there are few methods be proposed to aim on it. Meanwhile, most of existing databases only concentrate on the anti-spoofing of different kinds of attacks and ignore the environmental changes in real world applications. In this paper, we build a new 3D mask anti-spoofing database with more variations to simulate the real world scenario. The proposed database contains 12 masks from two companies with different appearance quality. 7 Cameras from the stationary and mobile devices and 6 lighting settings that cover typical illumination conditions are also included. Therefore, each subject contains 42 (7 cameras ∗ 6 lightings) genuine and 42 mask sequences and the total size is 1008 videos. Through the benchmark experiments, directions of the future study are pointed out. We plan to release the database as an platform to evaluate methods under different variations.
AB - 3D Mask face spoofing attack becomes new challenge and attracts more research interests in recent years. However, due to the deficiency number and limited variations of database, there are few methods be proposed to aim on it. Meanwhile, most of existing databases only concentrate on the anti-spoofing of different kinds of attacks and ignore the environmental changes in real world applications. In this paper, we build a new 3D mask anti-spoofing database with more variations to simulate the real world scenario. The proposed database contains 12 masks from two companies with different appearance quality. 7 Cameras from the stationary and mobile devices and 6 lighting settings that cover typical illumination conditions are also included. Therefore, each subject contains 42 (7 cameras ∗ 6 lightings) genuine and 42 mask sequences and the total size is 1008 videos. Through the benchmark experiments, directions of the future study are pointed out. We plan to release the database as an platform to evaluate methods under different variations.
UR - http://www.scopus.com/inward/record.url?scp=85010205084&partnerID=8YFLogxK
U2 - 10.1109/CVPRW.2016.193
DO - 10.1109/CVPRW.2016.193
M3 - Conference proceeding
AN - SCOPUS:85010205084
T3 - IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
SP - 1551
EP - 1557
BT - Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2016
PB - IEEE Computer Society
T2 - 29th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2016
Y2 - 26 June 2016 through 1 July 2016
ER -