Exploring Adversarial Fake Images on Face Manifold
This work addresses the vulnerability of deepfake detection models to adversarial attacks, which is a significant concern for digital forensics and security practitioners.
This paper explores generating anti-forensic fake face images by optimally searching adversarial points on the face manifold, rather than adding adversarial noise. The generated fake images, created by iteratively performing gradient descent in the latent space of a generative model like StyleGAN, can reduce the accuracy of deepfake detection models (e.g., Xception or EfficientNet) from over 90% to nearly 0% while maintaining high visual quality.
Images synthesized by powerful generative adversarial network (GAN) based methods have drawn moral and privacy concerns. Although image forensic models have reached great performance in detecting fake images from real ones, these models can be easily fooled with a simple adversarial attack. But, the noise adding adversarial samples are also arousing suspicion. In this paper, instead of adding adversarial noise, we optimally search adversarial points on face manifold to generate anti-forensic fake face images. We iteratively do a gradient-descent with each small step in the latent space of a generative model, e.g. Style-GAN, to find an adversarial latent vector, which is similar to norm-based adversarial attack but in latent space. Then, the generated fake images driven by the adversarial latent vectors with the help of GANs can defeat main-stream forensic models. For examples, they make the accuracy of deepfake detection models based on Xception or EfficientNet drop from over 90% to nearly 0%, meanwhile maintaining high visual quality. In addition, we find manipulating style vector $z$ or noise vectors $n$ at different levels have impacts on attack success rate. The generated adversarial images mainly have facial texture or face attributes changing.