Research map / 3D vision / NeRF and radiance fields
NVFi: Neural Velocity Fields for 3D Physics Learning from Dynamic Videos
Jinxi Li, Ziyang Song, Bo Yang
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 |
|---|---|---|---|
| Jointly optimize keyframe radiance and interframe velocity fields (ablation) | NVFi · Dynamic Object Dataset extrapolation | PSNR: 24.792 -> 27.594 | Improved |
| Add divergence-free and momentum physics constraints (ablation) | NVFi · Dynamic Object Dataset extrapolation | PSNR: 25.537 -> 27.594 | Improved |
| Use 8 keyframes rather than 16 (ablation) | NVFi · Dynamic Object Dataset extrapolation | PSNR: 27.594 -> 27.490 | No clear effect |
| Increase keyframes from 16 to 32 (ablation) | NVFi · Dynamic Object Dataset extrapolation | PSNR: 27.594 -> 24.902 | Hurt |
| Increase keyframes from 4 to 8 (ablation) | NVFi · Dynamic Indoor Scene Dataset extrapolation | PSNR: 29.745 -> 29.093 | Hurt |
| Increase keyframes from 4 to 16 (ablation) | NVFi · Dynamic Indoor Scene Dataset extrapolation | PSNR: 29.745 -> 26.235 | Hurt |
| Reduce training cameras from 12 to 6 (ablation) | NVFi · Dynamic Object Dataset extrapolation | PSNR: 27.594 -> 25.114 | Hurt |
| Reduce training cameras from 12 to 3 (ablation) | NVFi · Dynamic Object Dataset extrapolation | PSNR: 27.594 -> 21.370 | Hurt |
| Learn velocity with NVFi rather than D-NeRF deformation | Motion transfer to Gnome-new · Gnome-new motion transfer | PSNR: 16.124 -> 16.178 | No clear effect |
Later papers that compared with this method
3 later accepted papers report NVFi as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- ODE-GS: Latent ODEs for Dynamic Scene Extrapolation with 3D Gaussian Splatting (ICLR 2026 · better on 3 of 3 shared results)
- GrowFields: Compositional 4D Neural Fields for Topology-Changing Plant Growth (ECCV 2026 · better on 2 of 2 shared results)
- TRACE: Learning 3D Gaussian Physical Dynamics from Multi-view Videos (ICCV 2025 · better on 6 of 6 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- D-NeRF — D-NeRF: Neural Radiance Fields for Dynamic Scenes (CVPR 2021) · better on 1 of 1 shared result
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
- FreeGave: 3D Physics Learning from Dynamic Videos by Gaussian Velocity (CVPR 2025 · similarity 0.89)
- TFS-NeRF: Template-Free NeRF for Semantic 3D Reconstruction of Dynamic Scene (NeurIPS 2024 · similarity 0.88)
- OSN: Infinite Representations of Dynamic 3D Scenes from Monocular Videos (ICML 2024 · similarity 0.88)
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.87)
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.87)
- Joint Optimization of Neural Radiance Fields and Continuous Camera Motion from a Monocular Video (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.
