FakeFormer: Efficient Vulnerability-Driven Transformers for Generalisable Deepfake Detection
This work addresses the challenge of detecting facial forgeries for security and media integrity, offering an incremental improvement by adapting ViTs to better model localized artifacts.
The paper tackles the problem of deepfake detection by addressing the suboptimal performance of Vision Transformers (ViTs) compared to CNNs, proposing FakeFormer which enforces extraction of subtle inconsistency-prone information and achieves state-of-the-art results in generalization and computational cost across multiple datasets.
Recently, Vision Transformers (ViTs) have achieved unprecedented effectiveness in the general domain of image classification. Nonetheless, these models remain underexplored in the field of deepfake detection, given their lower performance as compared to Convolution Neural Networks (CNNs) in that specific context. In this paper, we start by investigating why plain ViT architectures exhibit a suboptimal performance when dealing with the detection of facial forgeries. Our analysis reveals that, as compared to CNNs, ViT struggles to model localized forgery artifacts that typically characterize deepfakes. Based on this observation, we propose a deepfake detection framework called FakeFormer, which extends ViTs to enforce the extraction of subtle inconsistency-prone information. For that purpose, an explicit attention learning guided by artifact-vulnerable patches and tailored to ViTs is introduced. Extensive experiments are conducted on diverse well-known datasets, including FF++, Celeb-DF, WildDeepfake, DFD, DFDCP, and DFDC. The results show that FakeFormer outperforms the state-of-the-art in terms of generalization and computational cost, without the need for large-scale training datasets. The code is available at \url{https://github.com/10Ring/FakeFormer}.