discovered 03 Aug 2026
FaceForensics
→ View on GitHubFaceForensics++ is a comprehensive dataset designed for detecting manipulated facial images, featuring 1000 original video sequences enhanced with four advanced face manipulation techniques: Deepfakes, Face2Face, FaceSwap, and NeuralTextures. It offers key features such as binary masks for image classification and segmentation tasks, an additional 1000 Deepfake models for data augmentation, and includes the FaceShifter method for high-fidelity identity-preserving face swaps, along with access to a benchmark for evaluating detection algorithms.