S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video.
Hao ZhangFang LiSamyak RawlekarNarendra AhujaPublished in: CoRR (2024)
Keyphrases
- articulated objects
- articulated human motion
- articulated motion
- shape model
- degrees of freedom
- human motion
- human body
- motion capture
- image sequences
- motion analysis
- rigid objects
- video sequences
- video data
- human pose
- optical flow
- single view
- pose estimation
- kinematic chain
- dynamic textures
- video frames
- rigid parts
- shape descriptors
- three dimensional
- human actions