Self-supervised Keypoint Correspondences for Multi-person Pose Estimation and Tracking in Videos.
Umer RafiAndreas DoeringBastian LeibeJuergen GallPublished in: ECCV (20) (2020)
Keyphrases
- human actions
- human pose
- pose estimation
- action recognition
- camera pose
- markerless
- monocular images
- computer vision
- video sequences
- human pose estimation
- body parts
- line correspondences
- static images
- tracking and pose estimation
- pose parameters
- articulated objects
- upper body
- pose recovery
- pose estimates
- motion tracking
- human body
- camera pose estimation
- d objects
- articulated model
- keypoints
- feature points
- image features
- kalman filter
- object recognition
- motion analysis
- head pose estimation
- degrees of freedom
- particle filter
- position and orientation
- focal length
- moving camera
- moving objects
- pictorial structure
- geometric constraints
- dynamic scenes
- video frames
- video surveillance
- visual tracking