Disentangled Visual Representations for Extreme Human Body Video Compression.
Ruofan WangQi MaoShiqi WangChuanmin JiaRonggang WangSiwei MaPublished in: ICME (2022)
Keyphrases
- video compression
- human body
- visual representations
- motion compensation
- motion estimation
- pedestrian detection
- video coding
- visual representation
- video data
- pose estimation
- visualization tools
- motion compensated
- body parts
- three dimensional
- human motion
- compression ratio
- concept maps
- video frames
- computer vision
- motion vectors
- image classification