Face Presentation Attack Detection by Excavating Causal Clues and Adapting Embedding Statistics

TL;DR

This paper introduces a novel face presentation attack detection method that leverages causal inference and feature distribution augmentation, achieving high effectiveness and efficiency without complex models or additional data.

Contribution

It models face PAD as a causal domain generalization task, excavates causal factors via counterfactuals, and employs class-guided MixStyle for data augmentation, all with minimal computational overhead.

Findings

Outperforms state-of-the-art PAD methods in cross-dataset tests.

Requires no extra trainable parameters or complex architectures.

Demonstrates high efficiency and effectiveness in experiments.

Abstract

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.