A Deepfake Face Image Detection Model Supporting Privacy Protection

Existing researches on deepfake face detection are all performed under plaintext conditions, while face images have significant privacy. Therefore, research on deepfake face image detection supporting privacy protection features is of great importance. This paper proposes a deepfake detection model supporting privacy protection for the first time. The model adopts the framework of additive secret sharing. Firstly, three secure communica-tion protocols are constructed on the basis of the existing fundamental computing protocols, and then a non-colluding dual server is used to construct a plaintext-like environment: with the support of the constructed secure communication protocols, the pre-trained ResNet50 models in dual servers compute interactively and cooperatively, and finally achieve secure deepfake face detection without exposing the in-puts. Experiments on the public available FaceForensics++ dataset show that the proposed security detection model can achieve the same accuracy as the ResNet50 model in plaintext while supporting privacy protection.