Skip to main content
GameDev.net gamedev.net
2 sources covering this story

DC-Motion: Decoupling Structure and Details via Discrete-Continuous Tokens for Human Motion Generation

A new motion-generation paper splits human movement into discrete structure and continuous detail, which is a useful idea for animation pipelines. The authors say this avoids the usual tradeoff between diffusion models’ smoothness and tokenized methods’ controllability. On HumanML3D and KIT-ML, it reportedly improves both FID and R-Precision.

First reported 3 months, 2 weeks ago • human-motion-generation animation text-to-motion diffusion-models
Want to discuss what this means for developers?
Open discussion

All Sources

View article page
PRIMARY SOURCE
arXiv cs.GR arXiv cs.GR

DC-Motion: Decoupling Structure and Details via Discrete-Continuous Tokens for Human Motion Generation

3 months, 2 weeks ago Read source
arXiv cs.GR arXiv cs.GR 1% match

Toward Richer Material Generation via Procedural Data Enhancement

3 months, 2 weeks ago Read source
Read Full Article Back to Feed