TY - GEN
T1 - Shuffled Patch-Wise Supervision for Presentation Attack Detection
AU - Kantarcı, Alperen
AU - Dertli, Hasan
AU - Ekenel, Hazım Kemal
N1 - Publisher Copyright:
© 2021 Gesellschaft fur Informatik (GI). All rights reserved.
PY - 2021
Y1 - 2021
N2 - Face anti-spoofing is essential to prevent false facial verification by using a photo, video, mask, or a different substitute for an authorized person’s face. Most of the state-of-the-art presentation attack detection (PAD) systems suffer from overfitting, where they achieve near-perfect scores on a single dataset but fail on a different dataset with more realistic data. This problem drives researchers to develop models that perform well under real-world conditions. This is an especially challenging problem for frame-based presentation attack detection systems that use convolutional neural networks (CNN). To this end, we propose a new PAD approach, which combines pixel-wise binary supervision with patch-based CNN. We believe that training a CNN with face patches allows the model to distinguish spoofs without learning background or dataset-specific traces. We tested the proposed method both on the standard benchmark datasets —Replay-Mobile, OULU-NPU—and on a real-world dataset. The proposed approach shows its superiority on challenging experimental setups. Namely, it achieves higher performance on OULU-NPU protocol 3, 4 and on inter-dataset real-world experiments.
AB - Face anti-spoofing is essential to prevent false facial verification by using a photo, video, mask, or a different substitute for an authorized person’s face. Most of the state-of-the-art presentation attack detection (PAD) systems suffer from overfitting, where they achieve near-perfect scores on a single dataset but fail on a different dataset with more realistic data. This problem drives researchers to develop models that perform well under real-world conditions. This is an especially challenging problem for frame-based presentation attack detection systems that use convolutional neural networks (CNN). To this end, we propose a new PAD approach, which combines pixel-wise binary supervision with patch-based CNN. We believe that training a CNN with face patches allows the model to distinguish spoofs without learning background or dataset-specific traces. We tested the proposed method both on the standard benchmark datasets —Replay-Mobile, OULU-NPU—and on a real-world dataset. The proposed approach shows its superiority on challenging experimental setups. Namely, it achieves higher performance on OULU-NPU protocol 3, 4 and on inter-dataset real-world experiments.
KW - Convolutional neural networks
KW - Face antispoofing
KW - Presentation attack detection
KW - Real-world dataset
UR - http://www.scopus.com/inward/record.url?scp=85127446686&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85127446686&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85127446686
T3 - Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)
SP - 61
EP - 69
BT - BIOSIG 2021 - Proceedings of the 20th International Conference of the Biometrics Special Interest Group
A2 - Bromme, Arslan
A2 - Busch, Christoph
A2 - Damer, Naser
A2 - Dantcheva, Antitza
A2 - Gomez-Barrero, Marta
A2 - Raja, Kiran
A2 - Rathgeb, Christian
A2 - Sequeira, Ana F.
A2 - Uhl, Andreas
PB - Gesellschaft fur Informatik (GI)
T2 - 20th International Conference of the Biometrics Special Interest Group, BIOSIG 2021
Y2 - 15 September 2021 through 17 September 2021
ER -