Face Presentation Attack Detection by Excavating Causal Clues and Adapting Embedding Statistics
This addresses performance degradation in face anti-spoofing systems when encountering unknown attack types or domains, though it appears incremental within existing domain generalization approaches.
The paper tackles face presentation attack detection (PAD) by modeling it as a compound domain generalization task from a causal perspective, using counterfactual intervention and class-guided MixStyle to improve performance without extra computational cost. Results show effectiveness and efficiency compared to state-of-the-art methods in cross-dataset experiments.
Recent face presentation attack detection (PAD) leverages domain adaptation (DA) and domain generalization (DG) techniques to address performance degradation on unknown domains. However, DA-based PAD methods require access to unlabeled target data, while most DG-based PAD solutions rely on a priori, i.e., known domain labels. Moreover, most DA-/DG-based methods are computationally intensive, demanding complex model architectures and/or multi-stage training processes. This paper proposes to model face PAD as a compound DG task from a causal perspective, linking it to model optimization. We excavate the causal factors hidden in the high-level representation via counterfactual intervention. Moreover, we introduce a class-guided MixStyle to enrich feature-level data distribution within classes instead of focusing on domain information. Both class-guided MixStyle and counterfactual intervention components introduce no extra trainable parameters and negligible computational resources. Extensive cross-dataset and analytic experiments demonstrate the effectiveness and efficiency of our method compared to state-of-the-art PADs. The implementation and the trained weights are publicly available.