InfoDense: Density-Aware Regional Decisive Replay for Memory-Efficient Incremental Face Forgery Detection

2026-07-18

AI Summary

This work addresses catastrophic forgetting, loss of forgery cues, and domain bias in incremental face forgery detection under limited memory constraints by proposing a density-aware regional key replay strategy. The method introduces a novel approach that leverages CLIP embeddings to localize image regions rich in forgery traces and selects samples based on both representativeness in the latent space and the number of critical patches. It balances information density and diversity through a three-stage pipeline of cropping, selection, and fusion, while an adaptive fusion mechanism mitigates replay-induced bias. Experimental results demonstrate that the proposed approach significantly alleviates forgetting, enhances cross-domain generalization, and substantially reduces memory overhead across multiple incremental deepfake benchmarks.

Abstract

The rapid evolution of face forgery techniques has introduced an increasing variety of manipulations. Incremental Face Forgery Detection (IFFD), which incrementally adds new forgery data to fine-tune previously trained models, has emerged as a promising approach to handle evolving forgery threats. However, conventional replay-based IFFD methods suffer from catastrophic forgetting. Storing full historical images under limited memory often either fails to preserve subtle forgery cues or introduces domain bias, reducing the model's ability to learn intrinsic and transferable manipulation characteristics. In this paper, we propose a Density-Aware Regional Decisive replay strategy, termed InfoDense, to address these challenges. InfoDense prioritizes artifact-dense and forgery-critical regions, significantly reducing storage requirements while maintaining high-fidelity forgery evidence. We first introduce InfoDense Cut to localize decisive patches using CLIP-based embeddings. Then, InfoDense Select ranks candidate segments by combining latent-space representativeness and decisive patch counts, ensuring both diversity and information density in the replay buffer. Finally, InfoDense Fuse reconstructs unbiased training inputs by adaptively merging stored segments with current-task samples, enhancing knowledge retention and generalization. Extensive experiments on challenging incremental deepfake benchmarks demonstrate that InfoDense effectively mitigates catastrophic forgetting while improving cross-domain generalization.

Problem

• Incremental Face Forgery Detection
• Catastrophic Forgetting
• Memory Efficiency
• Face Forgery Detection
• Replay-based Learning

Innovation

• Incremental Learning
• Face Forgery Detection
• Replay Strategy
• Information Density
• Catastrophic Forgetting