Direct Discrepancy Replay: Distribution-Discrepancy Condensation and Manifold-Consistent Replay for Continual Face Forgery Detection

2026-04-14

AI Summary

This work addresses the dual challenges of catastrophic forgetting and privacy leakage in continual face forgery detection, where models tend to forget previously learned forgery types when adapting to new ones, and existing replay-based methods struggle to maintain sample diversity while preserving privacy under stringent memory constraints. To overcome these limitations, the authors propose Distribution Discrepancy Compression (DDC) and Manifold-Consistent Replay (MCR), which, for the first time, focus replay on modeling the discrepancy between real and forged data distributions. By leveraging proxy decomposition in a characteristic function space and variance-preserving image synthesis, the method generates historical forgery samples without storing raw data. Evaluated under extremely limited memory budgets, the approach significantly outperforms current state-of-the-art methods, effectively mitigating catastrophic forgetting while reducing the risk of identity leakage.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Privacy
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Continual face forgery detection (CFFD) requires detectors to learn emerging forgery paradigms without forgetting previously seen manipulations. Existing CFFD methods commonly rely on replaying a small amount of past data to mitigate forgetting. Such replay is typically implemented either by storing a few historical samples or by synthesizing pseudo-forgeries from detector-dependent perturbations. Under strict memory budgets, the former cannot adequately cover diverse forgery cues and may expose facial identities, while the latter remains strongly tied to past decision boundaries. We argue that the core role of replay in CFFD is to reinstate the distributions of previous forgery tasks during subsequent training. To this end, we directly condense the discrepancy between real and fake distributions and leverage real faces from the current stage to perform distribution-level replay. Specifically, we introduce Distribution-Discrepancy Condensation (DDC), which models the real-to-fake discrepancy via a surrogate factorization in characteristic-function space and condenses it into a tiny bank of distribution discrepancy maps. We further propose Manifold-Consistent Replay (MCR), which synthesizes replay samples through variance-preserving composition of these maps with current-stage real faces, yielding samples that reflect previous-task forgery cues while remaining compatible with current real-face statistics. Operating under an extremely small memory budget and without directly storing raw historical face images, our framework consistently outperforms prior CFFD baselines and significantly mitigates catastrophic forgetting. Replay-level privacy analysis further suggests reduced identity leakage risk relative to selection-based replay.

Problem

• Continual Face Forgery Detection
• Catastrophic Forgetting
• Memory Budget
• Privacy Leakage
• Distribution Discrepancy

Innovation

• Distribution-Discrepancy Condensation
• Manifold-Consistent Replay
• Continual Face Forgery Detection
• Distribution-Level Replay
• Catastrophic Forgetting Mitigation