Search papers, labs, and topics across Lattice.
The paper identifies a critical flaw in cross-dataset evaluation of face forgery detectors, where standard AUC metrics fail to capture score distribution shifts across domains. To address this, they introduce Cross-AUC, a new metric that explicitly evaluates cross-domain score comparability by contrasting real and fake samples across different datasets. They then propose SFAM, a novel detection framework using patch-level image-text alignment and a mixture-of-experts module, which achieves state-of-the-art performance under both standard and the new Cross-AUC metrics.
Face forgery detectors crumble when evaluated on unseen data, but a new metric, Cross-AUC, finally exposes this hidden vulnerability.
Nowadays, visual data forgery detection plays an increasingly important role in social and economic security with the rapid development of generative models. Existing face forgery detectors still can't achieve satisfactory performance because of poor generalization ability across datasets. The key factor that led to this phenomenon is the lack of suitable metrics: the commonly used cross-dataset AUC metric fails to reveal an important issue where detection scores may shift significantly across data domains. To explicitly evaluate cross-domain score comparability, we propose \textbf{Cross-AUC}, an evaluation metric that can compute AUC across dataset pairs by contrasting real samples from one dataset with fake samples from another (and vice versa). It is interesting to find that evaluating representative detectors under the Cross-AUC metric reveals substantial performance drops, exposing an overlooked robustness problem. Besides, we also propose the novel framework \textbf{S}emantic \textbf{F}ine-grained \textbf{A}lignment and \textbf{M}ixture-of-Experts (\textbf{SFAM}), consisting of a patch-level image-text alignment module that enhances CLIP's sensitivity to manipulation artifacts, and the facial region mixture-of-experts module, which routes features from different facial regions to specialized experts for region-aware forgery analysis. Extensive qualitative and quantitative experiments on the public datasets prove that the proposed method achieves superior performance compared with the state-of-the-art methods with various suitable metrics.
