Physics-based motion capture imitation with deep reinforcement learning.
Nuttapong ChentanezMatthias MüllerMiles MacklinViktor MakoviychukStefan JeschkePublished in: MIG (2018)
Keyphrases
- motion capture
- reinforcement learning
- human body
- human motion
- human figure
- computer graphics
- human computer interaction
- human motion capture
- pose tracking
- state space
- monocular video sequences
- computer animation
- motion capture data
- machine learning
- human movement
- optimal policy
- multi view
- markerless
- body movements
- multi modal
- deformable models
- articulated human motion
- non rigid structure from motion
- human motion capture data
- dynamic programming
- human pose
- humanoid robot
- body parts
- transfer learning
- mobile robot
- moving objects