Research map / Video / Video generation
NeMF: Neural Motion Fields for Kinematic Animation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Represent the entire motion as a continuous time-conditioned field with a learned style latent | NeMF · AMASS motion reconstruction and synthesis | MRE: 13.008 -> 5.988 | Helped |
| Represent the entire motion as a continuous time-conditioned field with a learned style latent | NeMF · AMASS motion reconstruction and synthesis | FID: 8.687 -> 6.508 | Helped |
| Represent the entire motion as a continuous time-conditioned field with a learned style latent | NeMF · AMASS motion reconstruction and synthesis | MTE: 23.882 -> 29.692 | Hurt |
| Use NeMF for motion clips in-betweening | NeMF · Motion clips in-betweening, 10-, 20-, and 30-frame gaps | FID: 0.027 -> 0.024; 0.170 -> 0.141; 0.455 -> 0.365 | Helped |
| Use NeMF for motion clips in-betweening | NeMF · Motion clips in-betweening, 10-, 20-, and 30-frame gaps | Foot Skating: 1.193 -> 0.646; 1.200 -> 0.697; 1.023 -> 0.660 | Helped |
| Use NeMF for sparse keyframe in-betweening | NeMF · Sparse keyframe in-betweening, 5-frame interval | FID: 0.032 -> 0.085 | Hurt |
| Use NeMF for sparse keyframe in-betweening | NeMF · Sparse keyframe in-betweening, 10-, 15-, and 20-frame intervals | FID: 0.305 -> 0.302; 0.713 -> 0.612; 1.191 -> 0.879 | Helped |
| Add positional encoding of time (ablation) | NeMF · AMASS motion reconstruction and synthesis | MRE: 6.413 -> 5.988 | Helped |
| Add positional encoding of time (ablation) | NeMF · AMASS motion reconstruction and synthesis | FID: 2.146 -> 2.073 | Helped |
| Use separate encoders for local motion and root orientation (ablation) | NeMF · AMASS motion reconstruction and synthesis | MRE: 8.661 -> 5.988 | Helped |
| Use separate encoders for local motion and root orientation (ablation) | NeMF · AMASS motion reconstruction and synthesis | MOE: 5.982 -> 6.157 | Hurt |
| Use separate encoders for local motion and root orientation (ablation) | NeMF · AMASS motion reconstruction and synthesis | MTE: 35.391 -> 29.692 | Helped |
Closest papers by meaning
- NIFTY: Neural Object Interaction Fields for Guided Human Motion Synthesis (CVPR 2024 · similarity 0.86)
- Degrees of Freedom Matter: Inferring Dynamics from Point Trajectories (CVPR 2024 · similarity 0.85)
- How to Train Neural Field Representations: A Comprehensive Study and Benchmark (CVPR 2024 · similarity 0.85)
- Universal Humanoid Motion Representations for Physics-Based Control (ICLR 2024 · similarity 0.85)
- Object-centric 3D Motion Field for Robot Learning from Human Videos (NeurIPS 2025 · similarity 0.85)
- NRDF: Neural Riemannian Distance Fields for Learning Articulated Pose Priors (CVPR 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.