Research map / Video / Video generation
Lagrangian Motion Fields for Long-Term Motion Generation
Yifei Yang, Zikai Huang, Chenshu Xu, Shengfeng He
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 |
|---|---|---|---|
| replace per-frame motion with supermotion representation | EDGE · FineDance | runtime: 8.59s -> 2.48s | Improved |
| replace per-frame motion with supermotion representation | EDGE · FineDance | FID_g: 50.38 -> 31.63 | Improved |
| replace per-frame motion with supermotion representation | EDGE · FineDance | FID_k: 94.34 -> 50.27 | Improved |
| replace per-frame motion with supermotion representation | Lodge (DDIM) · FineDance | FID_k: 50.00 -> 52.98 | Hurt |
| replace per-frame motion with supermotion representation | Lodge (DDIM) · FineDance | FID_g: 35.52 -> 23.39 | Improved |
| use supermotion instead of Bailando with K-Means latent codes | Long-term Bailando + K-Means · FineDance | FID_k: 72.52 -> 50.27 | Improved |
| replace per-frame motion with supermotion representation | MDM · HumanML3D | R-Precision top-3: 0.611 -> 0.737 | Improved |
| replace per-frame motion with supermotion representation | MDM · HumanML3D | FID: 0.544 -> 1.736 | Hurt |
| replace per-frame motion with supermotion representation | MDM · HumanML3D-MP | FID: 2.315 -> 1.668 | Improved |
| replace VQ-VAE latent with supermotion | T2M-GPT · HumanML3D-MP | FID: 2.083 -> 1.586 | Improved |
| add the coherent loss (ablation) | Ours (3D) · FineDance | FID_g: 36.37 -> 33.84 | Improved |
| add the coherent loss (ablation) | Ours (3D) · HumanML3D | FID: 1.760 -> 1.750 | Improved |
Closest papers by meaning
- MotionMix: Weakly-Supervised Diffusion for Controllable Motion Generation (AAAI 2024 · similarity 0.85)
- FLD: Fourier Latent Dynamics for Structured Motion Representation and Learning (ICLR 2024 · similarity 0.85)
- Generative Image Dynamics (CVPR 2024 · similarity 0.85)
- Deep Compositional Phase Diffusion for Long Motion Sequence Generation (NeurIPS 2025 · similarity 0.85)
- PhysCtrl: Generative Physics for Controllable and Physics-Grounded Video Generation (NeurIPS 2025 · similarity 0.85)
- Human Motion Diffusion as a Generative Prior (ICLR 2024 · similarity 0.84)
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.
