DFIL: Deepfake Incremental Learning by Exploiting Domain-invariant Forgery Clues
This addresses the trust crisis caused by deepfakes by enhancing detection models' adaptability to new forgery methods, though it is incremental as it builds on existing incremental learning techniques.
The paper tackles the problem of deepfake detection models degrading in accuracy on images from new generation methods by proposing an incremental learning framework that improves generalization using a small number of new samples, achieving a state-of-the-art average forgetting rate of 7.01 and average accuracy of 85.49 on benchmark datasets.
The malicious use and widespread dissemination of deepfake pose a significant crisis of trust. Current deepfake detection models can generally recognize forgery images by training on a large dataset. However, the accuracy of detection models degrades significantly on images generated by new deepfake methods due to the difference in data distribution. To tackle this issue, we present a novel incremental learning framework that improves the generalization of deepfake detection models by continual learning from a small number of new samples. To cope with different data distributions, we propose to learn a domain-invariant representation based on supervised contrastive learning, preventing overfit to the insufficient new data. To mitigate catastrophic forgetting, we regularize our model in both feature-level and label-level based on a multi-perspective knowledge distillation approach. Finally, we propose to select both central and hard representative samples to update the replay set, which is beneficial for both domain-invariant representation learning and rehearsal-based knowledge preserving. We conduct extensive experiments on four benchmark datasets, obtaining the new state-of-the-art average forgetting rate of 7.01 and average accuracy of 85.49 on FF++, DFDC-P, DFD, and CDF2. Our code is released at https://github.com/DeepFakeIL/DFIL.