TY - GEN
T1 - Temporal similarity analysis of remote photoplethysmography for fast 3D mask face presentation attack detection
AU - Liu, Si Qi
AU - Lan, Xiangyuan
AU - Yuen, Pong Chi
N1 - Funding Information:
Besides, we can find also that compared with the face resolution and lighting condition, camera with different compression settings affects more on the performance. For instance, although HKBU-MARSv1+ is recorded at high resolution compared with 3DMAD and HKBU-MARsV2+, the performances are lower (see Tab. 1). This is because the H.264 compression removes some of the subtle color variations that reflect heartbeat [28]. More self-made 3D mask attack datasets with different camera and compression settings are needed to further investigate the properties of rPPG-based 3D mask PAD. 7. Acknowledgement This project is partially supported by Hong Kong RGC General Research Fund HKBU 12201215 and HKBU Tier 1 start-up Grant.
PY - 2020/3
Y1 - 2020/3
N2 - To tackle the 3D mask face presentation attack, remote Photoplethysmography (rPPG), a biomedical technique that can detect heartbeat signal remotely, is employed as an intrinsic liveness cue. Although existing rPPG-based methods exhibit encouraging results, they require long observation time (10-12 seconds) to identify the heartbeat information, which limits their employment in real applications such as smartphone unlock and e-payment. To shorten the observation time (within 1-second) while keeping the performance, we propose a fast rPPG-based 3D mask presentation attack detection (PAD) method by analyzing the similarity of local facial rPPG signals in the time domain. In particular, a set of temporal similarity features of facial and background local rPPG signals are designed and fused to adapt the real world variations based on rPPG shape and phase properties. For better evaluation under practical variations, we build the HKBU-MARsV2+ dataset that includes 16 masks from 2 types and 6 lighting conditions. Finally, extensive experiments are conducted on 11092 shortterm video slots from 4 datasets with a large number of real- world variations, in terms of mask type, lighting condition, camera, resolution of face region, and compression setting. Results show that the proposed TSrPPG outperforms the state-of-the-art competitors dramatically on discriminabil- ity and generalizability. To our best knowledge, this is the first work that addresses the length of observation time issue of rPPG-based 3D mask PAD.
AB - To tackle the 3D mask face presentation attack, remote Photoplethysmography (rPPG), a biomedical technique that can detect heartbeat signal remotely, is employed as an intrinsic liveness cue. Although existing rPPG-based methods exhibit encouraging results, they require long observation time (10-12 seconds) to identify the heartbeat information, which limits their employment in real applications such as smartphone unlock and e-payment. To shorten the observation time (within 1-second) while keeping the performance, we propose a fast rPPG-based 3D mask presentation attack detection (PAD) method by analyzing the similarity of local facial rPPG signals in the time domain. In particular, a set of temporal similarity features of facial and background local rPPG signals are designed and fused to adapt the real world variations based on rPPG shape and phase properties. For better evaluation under practical variations, we build the HKBU-MARsV2+ dataset that includes 16 masks from 2 types and 6 lighting conditions. Finally, extensive experiments are conducted on 11092 shortterm video slots from 4 datasets with a large number of real- world variations, in terms of mask type, lighting condition, camera, resolution of face region, and compression setting. Results show that the proposed TSrPPG outperforms the state-of-the-art competitors dramatically on discriminabil- ity and generalizability. To our best knowledge, this is the first work that addresses the length of observation time issue of rPPG-based 3D mask PAD.
UR - http://www.scopus.com/inward/record.url?scp=85085473202&partnerID=8YFLogxK
U2 - 10.1109/WACV45572.2020.9093337
DO - 10.1109/WACV45572.2020.9093337
M3 - Conference contribution
AN - SCOPUS:85085473202
T3 - Proceedings - 2020 IEEE Winter Conference on Applications of Computer Vision, WACV 2020
SP - 2597
EP - 2605
BT - Proceedings - 2020 IEEE Winter Conference on Applications of Computer Vision, WACV 2020
PB - IEEE
T2 - 2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020
Y2 - 1 March 2020 through 5 March 2020
ER -