Research map / Video / Video generation
Local Action-Guided Motion Diffusion Model for Text-to-Motion Generation
Peng Jin, Hao Li, Zesen Cheng, Kehan Li, Runyi Yu, Chang Liu, Xiangyang Ji, Li Yuan, Jie Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add local action guidance to the motion+action level model (ablation) | GuidedMotion · HumanML3D test | FID: 0.133 -> 0.119 | Improved |
| add local action guidance to the full three-level model (ablation) | GuidedMotion · HumanML3D test | FID: 0.107 -> 0.057 | Improved |
| add action level to motion-level-only model (ablation) | GuidedMotion · HumanML3D test | FID: 0.186 -> 0.133 | Improved |
| add specific level to model with motion+action levels, no guidance (ablation) | GuidedMotion · HumanML3D test | FID: 0.133 -> 0.107 | Improved |
| add specific level without guidance (ablation) | GuidedMotion · HumanML3D test | R-Precision Top-3: 0.771 -> 0.769 | Hurt |
| increase VAE token size from 2 to 8 (ablation) | motion VAE · HumanML3D test | FID: 1.906 -> 0.019 | Improved |
| increase VAE token size from 2 to 8 (ablation) | motion VAE · KIT test | R-Precision Top-3: 0.745 -> 0.760 | Mixed |
| allocate more diffusion steps per level (150 vs 300 total) (ablation) | GuidedMotion · HumanML3D test | FID: 0.057 vs 0.062 | No clear effect |
| use local-to-global paradigm instead of direct global generation | GuidedMotion · HumanML3D complex motion subset | FID: 0.314 -> 0.144 | Improved |
| split diffusion into three levels with same total steps | GuidedMotion vs MLD · HumanML3D test | inference time (s): 2.4998 -> 2.5189 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- ReMoDiffuse — ReMoDiffuse: Retrieval-Augmented Motion Diffusion Model (ICCV 2023) · better on 1 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Enhanced Fine-Grained Motion Diffusion for Text-Driven Human Motion Synthesis (AAAI 2024 · similarity 0.91)
- OMG: Towards Open-vocabulary Motion Generation via Mixture of Controllers (CVPR 2024 · similarity 0.89)
- Towards Detailed Text-to-Motion Synthesis via Basic-to-Advanced Hierarchical Diffusion Model (AAAI 2024 · similarity 0.89)
- Human Motion Diffusion as a Generative Prior (ICLR 2024 · similarity 0.89)
- AMD: Autoregressive Motion Diffusion (AAAI 2024 · similarity 0.89)
- MotionMix: Weakly-Supervised Diffusion for Controllable Motion Generation (AAAI 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
