Self-Supervised Equivariant Scene Synthesis from Video.
Cinjon ResnickOr LitanyCosmas HeißHugo LarochelleJoan BrunaKyunghyun ChoPublished in: CoRR (2021)
Keyphrases
- video sequences
- video images
- dynamic scenes
- scene change detection
- moving camera
- input video
- video scene
- space time
- multiple video streams
- live video
- video data
- d scene
- video frames
- visual data
- virtual view
- surveillance videos
- image frames
- single frame
- stationary camera
- real time
- multimedia
- image mosaics
- image sequences
- video analysis
- three dimensional
- scene analysis
- object detection and tracking
- video footage
- video streams
- complex scenes
- layered representation
- high resolution video
- input image
- multiple frames
- crowded scenes
- motion features
- video content
- object motion
- scene understanding
- object tracking
- single image
- video surveillance
- video retrieval
- motion estimation
- moving objects
- neighboring frames
- key frames
- temporal coherence
- dynamic textures
- real scenes
- foreground objects