HUMANISE: Language-conditioned Human Motion Generation in 3D Scenes.
Zan WangYixin ChenTengyu LiuYixin ZhuWei LiangSiyuan HuangPublished in: CoRR (2022)
Keyphrases
- human motion
- d scene
- human body
- motion capture
- video sequences
- image sequences
- optical motion capture
- image based rendering
- depth map
- single image
- motion sequences
- human actions
- motion recognition
- body parts
- optical flow
- human movements
- photorealistic
- spatio temporal
- moving camera
- motion capture data
- three dimensional
- high quality
- motion synthesis
- view dependent
- object tracking
- motion history images
- camera motion
- multi view