Rendering Humans from Object-Occluded Monocular Videos.
Tiange XiangAdam SunJiajun WuEhsan AdeliLi Fei-FeiPublished in: CoRR (2023)
Keyphrases
- object tracking
- partially occluded
- video sequences
- image sequences
- real time
- high quality
- high definition
- d objects
- human object interactions
- independently moving objects
- video content
- computer graphics
- pose estimation
- video surveillance
- occluded objects
- pose recovery
- object model
- moving objects
- human pose
- object recognition
- view invariant
- optical flow
- video data
- photorealistic
- video analysis
- multiple objects
- key frames
- appearance model
- video frames