DeepfakeMAE: Facial Part Consistency Aware Masked Autoencoder for Deepfake Video Detection.
Juan HuXin LiaoDifei GaoSatoshi TsutsuiZheng QinMike Zheng ShouPublished in: CoRR (2023)
Keyphrases
- landmark detection
- face detection and tracking
- video sequences
- activity detection
- video streams
- event detection
- video data
- false positives
- automatic detection
- video indexing
- detection method
- real time
- video content
- object detection and tracking
- object detection
- face recognition
- age estimation
- multimedia
- false alarms
- video analysis
- video frames
- facial expressions
- detection accuracy
- key frames
- video scene
- shot detection
- multi modal fusion
- facial images
- visual analysis
- temporal coherence
- temporal consistency
- text detection
- human faces
- detection algorithm
- face images