Research map / 3D vision / Gaussian splatting
PREF: Predictability Regularized Neural Motion Fields
Liangchen Song, Xuan Gong, Benjamin Planche, Meng Zheng, David Doermann, Junsong Yuan, Terrence Chen, Ziyan Wu
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 predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Sports | mMPJPE_5 (cm): 5.171 -> 4.533 | Improved |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Tools | mMPJPE_5 (cm): 2.341 -> 1.684 | Improved |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Cello | mMPJPE_10 (cm): 3.296 -> 2.572 | Improved |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Ian | mMPJPE_5 (cm): 2.349 -> 2.402 | Hurt |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Ian | mMPJPE_10 (cm): 3.534 -> 3.792 | Hurt |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Ian | mMPJPE_15 (cm): 5.282 -> 4.980 | Improved |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Sports | mMPJPE_15 (cm): 11.254 -> 10.718 | Improved |
| add predictability regularization (PREF) vs best baseline | neural motion field · Panoptic Tools | mMPJPE_15 (cm): 3.984 -> 3.393 | Improved |
| use shared embedding basis vectors instead of per-frame embeddings (ablation) | motion predictor · training loss | convergence speed | Improved |
Closest papers by meaning
- Degrees of Freedom Matter: Inferring Dynamics from Point Trajectories (CVPR 2024 · similarity 0.87)
- DynMF: Neural Motion Factorization for Real-time Dynamic View Synthesis with 3D Gaussian Splatting (ECCV 2024 · similarity 0.86)
- NVFi: Neural Velocity Fields for 3D Physics Learning from Dynamic Videos (NeurIPS 2023 · similarity 0.85)
- DIMO: Diverse 3D Motion Generation for Arbitrary Objects (ICCV 2025 · similarity 0.85)
- 3D Reconstruction with Generalizable Neural Fields using Scene Priors (ICLR 2024 · similarity 0.84)
- EmerNeRF: Emergent Spatial-Temporal Scene Decomposition via Self-Supervision (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.
