Research map / Video / Video generation
MotionLab: Unified Human Motion Generation and Editing via the Motion-Condition-Motion Paradigm
Ziyan Guo, Zeyu Hu, De Wen Soh, Na Zhao
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 motion curriculum learning (easy-to-hard task order with masked pre-training) (ablation) | MotionLab · HumanML3D text generation | FID: 1.956 -> 0.167 | Improved |
| add Aligned ROPE (1D) instead of 1D learnable encoding (ablation) | MotionLab · HumanML3D trajectory generation | Average Error: 0.0886 -> 0.0334 | Improved |
| use MotionFlow Transformer instead of regular cross-attention transformer (ablation) | MotionLab · HumanML3D text generation | FID: 0.483 -> 0.167 | Improved |
| use rectified flows instead of diffusion (ablation) | MotionLab · HumanML3D text generation | FID: 0.301 -> 0.167 | Improved |
| add Task Instruction Modulation with CLIP text instructions (ablation) | MotionLab · HumanML3D text generation | FID: 0.223 -> 0.167 | Improved |
| train unified multi-task model instead of per-task specialists (ablation) | MotionLab · MotionFix text-based editing | R@1: 41.44 -> 56.34 | Improved |
| use 1D Aligned ROPE instead of 3D ROPE (ablation) | MotionLab · HumanML3D text generation | FID: 0.241 -> 0.167 | Improved |
| add masked pre-training before supervised fine-tuning (ablation) | MotionLab · HumanML3D text generation | FID: 0.861 -> 0.167 | Improved |
| replace one-hot or learnable task tokens with text instructions (ablation) | MotionLab · HumanML3D text generation | FID: 0.187 -> 0.167 | Improved |
Later papers that compared with this method
1 later accepted paper reports MotionLab as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- InterEdit: Navigating Text-Guided Multi-Human 3D Motion Editing (ECCV 2026 · better on 7 of 7 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
- MMM: Generative Masked Motion Model (CVPR 2024 · similarity 0.88)
- Scaling Large Motion Models with Million-Level Human Motions (ICML 2025 · similarity 0.88)
- Move-in-2D: 2D-Conditioned Human Motion Generation (CVPR 2025 · similarity 0.88)
- MotionPro: A Precise Motion Controller for Image-to-Video Generation (CVPR 2025 · similarity 0.88)
- Motion Prompting: Controlling Video Generation with Motion Trajectories (CVPR 2025 · similarity 0.88)
- Dynamic Motion Blending for Versatile Motion Editing (CVPR 2025 · similarity 0.87)
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.
