PVASS-MDD: Predictive Visual-Audio Alignment Self-Supervision for Multimodal Deepfake Detection.
Yang YuXiaolong LiuRongrong NiSiyuan YangYao ZhaoAlex C. KotPublished in: IEEE Trans. Circuits Syst. Video Technol. (2024)
Keyphrases
- visual information
- cross modal
- multimodal information
- audio visual
- multi modal
- visual data
- multimedia
- detection method
- automatic detection
- false alarms
- visual features
- image alignment
- thermal images
- detection accuracy
- single modality
- low level
- false positives
- video data
- detection algorithm
- multimodal fusion
- multi stream
- video recordings
- soccer video
- visual analysis
- visual perception
- visual cues
- high level