LatentForensics: Towards lighter deepfake detection in the StyleGAN latent space

The classification of forged videos has been a challenge for the past few years. Deepfake classifiers can now reliably predict whether or not video frames have been tampered with. However, their performance is tied to both the dataset used for training and the analyst's computational power. We propose a deepfake classification method that operates in the latent space of a state-of-the-art generative adversarial network (GAN) trained on high-quality face images. The proposed method leverages the structure of the latent space of StyleGAN to learn a lightweight classification model. Experimental results on a standard dataset reveal that the proposed approach outperforms other state-of-the-art deepfake classification methods. To the best of our knowledge, this is the first study showing the interest of the latent space of StyleGAN for deepfake classification. Combined with other recent studies on the interpretation and manipulation of this latent space, we believe that the proposed approach can help in developing robust deepfake classification methods based on interpretable high-level properties of face images.