LGTM: Local-to-Global Text-Driven Human Motion Diffusion Model.
Haowen SunRuikun ZhengHaibin HuangChongyang MaHui HuangRuizhen HuPublished in: SIGGRAPH (Conference Paper Track) (2024)
Keyphrases
- human motion
- diffusion model
- motion capture
- human body
- optical motion capture
- video sequences
- information diffusion
- body parts
- anisotropic diffusion
- image sequences
- human actions
- spatio temporal
- motion sequences
- diffusion process
- motion recognition
- text mining
- motion capture data
- object recognition
- human movements
- motion synthesis
- motion estimation
- d objects
- visual data
- social media
- feature extraction