Flowpose: Conditional Normalizing Flows for 3D Human Pose and Shape Estimation from Monocular Videos.
Yaoyao DuZixiao ZhangZhihao LiPeng WeiQingmin LiaoWenming YangPublished in: ICASSP (2023)
Keyphrases
- shape estimation
- monocular images
- shape from shading
- shape modeling
- human pose
- pose estimation
- inverse kinematics
- structure from motion
- video frames
- inverse problems
- image sequences
- human body
- video sequences
- depth estimation
- human actions
- action recognition
- video data
- body parts
- real objects
- convex optimization
- optical flow