README.md

Overview

Title: One Detector to Rule Them All: Towards a General Deepfake Attack Detection Framework

Additional Results

Updated in-domain attack results including DFDC dataset

• Note that CLRNet performs the best for DFDC dataset among all the test baselines.

Updated out-of-domain attack results (before using our defense strategy)

• Note that results from Table 5 demonstrates that models trained on DFDC, which is a quite generic and diverse dataset, can still fail to detect out-of-domain attack (see Table 5).
• See Table 6 in our paper, for defense performance against out-of-domain attack.

Dataset used for Evaluation

• Facial Reenactment
  • Face2Face (F2F) [Dataset] [Paper]
  • Neural Texture (NT) [Dataset] [Paper]
• Identity Swap
  • DeepFake (DF) [Dataset] [GitHub]
  • FaceSwap (FS) [Dataset] [GitHub]
  • DeepFake Detection (DFD) [Dataset] [Blog]
  • DeepFake Detection Challenge (DFDC) [Dataset] [Paper] [Kaggle Challenge] [Blog]
• Unknown
  • DeepFake in the Wild (DFW) [Sample Links] [Source1] [Source2] [Source3]

Models used for Evaluation

• Single Frame-based
  • ShallowNet [Paper] [GitHub]
  • Xception [Paper] [GitHub]
  • MesoNet [Paper] [GitHub]
• Multiple Frame-based
  • CNN+LSTM [Paper]
  • DBiRNN [Paper]
  • CLRNet (Ours)