Enhancing Self-Supervised Talking Head Forgery Detection via a Training-Free Dual-System Framework

2026-05-05

AI Summary

This work addresses the limited discriminative capacity of existing self-supervised talking-head forgery detection methods on challenging samples, which leads to unreliable anomaly ranking and poor generalization. Inspired by the dual-process theory of human cognition, we propose a training-free dual-system framework (TFDS) that routes samples via a lightweight thresholding mechanism into confident and uncertain subsets. For the uncertain subset, TFDS introduces fine-grained evidence-guided re-ranking to recalibrate anomaly scores. To our knowledge, this is the first approach to incorporate dual-system cognitive mechanisms into forgery detection, effectively uncovering underutilized discriminative cues within existing detectors without incurring additional training costs. Extensive experiments demonstrate consistent performance gains across multiple datasets and perturbation settings, significantly outperforming baseline methods.

Abstract

Supervised talking head forgery detection faces severe generalization challenges due to the continuous evolution of generators. By reducing reliance on generator-specific forgery patterns, self-supervised detectors offer stronger cross-generator robustness. However, existing research has mainly focused on building stronger detectors, while the discriminative capacity of trained detectors remains insufficiently exploited. In particular, for score-based self-supervised detectors, the limited discriminative ability on hard cases is often reflected in unreliable anomaly ordering, leaving room for further refinement. Motivated by this observation, we draw inspiration from the dual-system theory of human cognition and propose a Training-Free Dual-System (TFDS) framework to further exploit the latent discriminative capacity of existing score-based self-supervised detectors. TFDS treats anomaly-like scores as the basis of System-1, using lightweight threshold-based routing to partition samples into confident and uncertain subsets. System-2 then revisits only the uncertain subset, performing fine-grained evidence-guided reasoning to refine the relative ordering of ambiguous samples within the original score distribution. Extensive experiments demonstrate consistent improvements across datasets and perturbation settings, with the gains arising mainly from corrected ordering within the uncertain subset. These findings show that existing self-supervised talking head forgery detectors still contain underexploited discriminative cues that can be effectively unlocked through training-free dual-system reasoning.

Problem

• talking head forgery detection
• self-supervised learning
• discriminative capacity
• anomaly ordering
• generalization

Innovation

• self-supervised detection
• talking head forgery
• dual-system reasoning
• training-free framework
• anomaly scoring