Video Autoencoder: self-supervised disentanglement of static 3D structure and motion.
Zihang LaiSifei LiuAlexei A. EfrosXiaolong WangPublished in: ICCV (2021)
Keyphrases
- space time
- spatial and temporal
- key frames
- temporal structure
- video data
- temporal filtering
- motion features
- video streams
- real time
- successive frames
- moving camera
- visual data
- image sequences
- temporal continuity
- spatio temporal
- geometric structure
- moving objects
- surveillance videos
- input video
- video sequences
- spatial and temporal information
- motion capture data
- low frame rate
- temporal consistency
- image frames
- dynamic textures
- motion capture
- motion analysis
- multimedia