datamllab/awesome-deepfakes-materials

Comprehensively catalogs deepfake resources across three modalities—videos/images (using GANs for generation and artifact-detection methods), synthetic text (leveraging Transformer-based language models), and voice synthesis—with separate sections for generation papers, detection approaches, implementation code, and benchmark datasets. The curation emphasizes both face-swapping techniques (FaceSwap, DeepFaceLab) and detection strategies spanning artifact analysis, binary classifiers, and cross-domain generalization. Includes linked research papers, open-source implementations, public datasets (FaceForensics++, Celeb-DF, DeepfakeTIMIT), and community challenges for evaluating detection robustness.

437 stars. No commits in the last 6 months.