Fooling State-of-the-Art Deepfake Detection with High-Quality Deepfakes
This addresses security and privacy threats from deepfakes by highlighting the need for better training data, though it is incremental as it builds on existing detection methods.
The paper tackles the problem of deepfake detectors failing in real-world scenarios by generating 90 high-quality deepfakes using a novel autoencoder and advanced blending technique, which drastically reduces a state-of-the-art detector's performance and provides useful clues for improvement when fine-tuned.
Due to the rising threat of deepfakes to security and privacy, it is most important to develop robust and reliable detectors. In this paper, we examine the need for high-quality samples in the training datasets of such detectors. Accordingly, we show that deepfake detectors proven to generalize well on multiple research datasets still struggle in real-world scenarios with well-crafted fakes. First, we propose a novel autoencoder for face swapping alongside an advanced face blending technique, which we utilize to generate 90 high-quality deepfakes. Second, we feed those fakes to a state-of-the-art detector, causing its performance to decrease drastically. Moreover, we fine-tune the detector on our fakes and demonstrate that they contain useful clues for the detection of manipulations. Overall, our results provide insights into the generalization of deepfake detectors and suggest that their training datasets should be complemented by high-quality fakes since training on mere research data is insufficient.