Towards Generalizable and Robust Face Manipulation Detection via Bag-of-local-feature
This addresses the issue of malicious facial manipulation abuse, offering improved detection for security applications, though it appears incremental as it builds on existing Transformer and bag-of-feature approaches.
The paper tackled the problem of poor generalization and robustness in face manipulation detection by proposing a bag-of-local-feature method, achieving state-of-the-art performance on datasets like FaceForensics++, Celeb-DF, and DeeperForensics-1.0.
Over the past several years, in order to solve the problem of malicious abuse of facial manipulation technology, face manipulation detection technology has obtained considerable attention and achieved remarkable progress. However, most existing methods have very impoverished generalization ability and robustness. In this paper, we propose a novel method for face manipulation detection, which can improve the generalization ability and robustness by bag-of-local-feature. Specifically, we extend Transformers using bag-of-feature approach to encode inter-patch relationships, allowing it to learn local forgery features without any explicit supervision. Extensive experiments demonstrate that our method can outperform competing state-of-the-art methods on FaceForensics++, Celeb-DF and DeeperForensics-1.0 datasets.