UCF: Uncovering Common Features for Generalizable Deepfake Detection
This addresses the challenge of generalizable deepfake detection for security and media integrity, though it is incremental as it builds on existing disentanglement and multi-task learning approaches.
The paper tackles the problem of deepfake detection generalizing poorly to new forgery types by addressing overfitting to irrelevant and method-specific features, proposing a disentanglement framework to uncover common forgery features, and achieves superior generalization compared to state-of-the-art methods.
Deepfake detection remains a challenging task due to the difficulty of generalizing to new types of forgeries. This problem primarily stems from the overfitting of existing detection methods to forgery-irrelevant features and method-specific patterns. The latter has been rarely studied and not well addressed by previous works. This paper presents a novel approach to address the two types of overfitting issues by uncovering common forgery features. Specifically, we first propose a disentanglement framework that decomposes image information into three distinct components: forgery-irrelevant, method-specific forgery, and common forgery features. To ensure the decoupling of method-specific and common forgery features, a multi-task learning strategy is employed, including a multi-class classification that predicts the category of the forgery method and a binary classification that distinguishes the real from the fake. Additionally, a conditional decoder is designed to utilize forgery features as a condition along with forgery-irrelevant features to generate reconstructed images. Furthermore, a contrastive regularization technique is proposed to encourage the disentanglement of the common and specific forgery features. Ultimately, we only utilize the common forgery features for the purpose of generalizable deepfake detection. Extensive evaluations demonstrate that our framework can perform superior generalization than current state-of-the-art methods.