Fusion methods for face presentation attack detection.

Abdullakutty, Faseela; Johnston, Pamela; Elyan, Eyad

Abstract

Face presentation attacks (PA) are a serious threat to face recognition (FR) applications. These attacks are easy to execute and difficult to detect. An attack can be carried out simply by presenting a video, photo, or mask to the camera. The literature shows that both modern, pre-trained, deep learning-based methods, and traditional hand-crafted, feature-engineered methods have been effective in detecting PAs. However, the question remains as to whether features learned in existing, deep neural networks sufficiently encompass traditional, low-level features in order to achieve optimal performance on PA detection tasks. In this paper, we present a simple feature-fusion method that integrates features extracted by using pre-trained, deep learning models with more traditional colour and texture features. Extensive experiments clearly show the benefit of enriching the feature space to improve detection rates by using three common public datasets, namely CASIA, Replay Attack, and SiW. This work opens future research to improve face presentation attack detection by exploring new characterizing features and fusion strategies.

Citation

ABDULLAKUTTY, F., JOHNSTON, P. and ELYAN, E. 2022. Fusion methods for face presentation attack detection. Sensors [online], 22(14): soft sensors 2021-2022, article 5196. Available from: https://doi.org/10.3390/s22145196

• Journal Article Type: Article
• Acceptance Date: Jul 7, 2022
• Online Publication Date: Jul 12, 2022
• Publication Date: Jul 31, 2022
• Deposit Date: Jul 22, 2022
• Publicly Available Date: Jul 22, 2022
• Journal: Sensors
• Print ISSN: 1424-8220
• Electronic ISSN: 1424-8220
• Publisher: MDPI
• Peer Reviewed: Peer Reviewed
• Volume: 22
• Issue: 14
• Article Number: 5196
• DOI: https://doi.org/10.3390/s22145196
• Keywords: Face presentation attacks; Deep learning; Feature-fusion; Spoofing detection; Liveness detection
• Public URL: https://rgu-repository.worktribe.com/output/1713269