DiffFAS: Face Anti-Spoofing via Generative Diffusion Models
This addresses the generalization challenge in FAS systems for security applications, though it is incremental as it builds on existing diffusion models for a specific domain.
The paper tackles the problem of domain shift in face anti-spoofing (FAS) by deconstructing it into image style and quality factors, proposing DiffFAS, a framework that uses generative diffusion models to transform live faces into high-fidelity attack faces, achieving state-of-the-art performance on cross-domain and cross-attack datasets.
Face anti-spoofing (FAS) plays a vital role in preventing face recognition (FR) systems from presentation attacks. Nowadays, FAS systems face the challenge of domain shift, impacting the generalization performance of existing FAS methods. In this paper, we rethink about the inherence of domain shift and deconstruct it into two factors: image style and image quality. Quality influences the purity of the presentation of spoof information, while style affects the manner in which spoof information is presented. Based on our analysis, we propose DiffFAS framework, which quantifies quality as prior information input into the network to counter image quality shift, and performs diffusion-based high-fidelity cross-domain and cross-attack types generation to counter image style shift. DiffFAS transforms easily collectible live faces into high-fidelity attack faces with precise labels while maintaining consistency between live and spoof face identities, which can also alleviate the scarcity of labeled data with novel type attacks faced by nowadays FAS system. We demonstrate the effectiveness of our framework on challenging cross-domain and cross-attack FAS datasets, achieving the state-of-the-art performance. Available at https://github.com/murphytju/DiffFAS.