Research map / 3D vision / Gaussian splatting
Shape of Motion: 4D Reconstruction from a Single Video
Qianqian Wang, Vickie Ye, Hang Gao, Weijia Zeng, Jake Austin, Zhengqi Li, Angjoo Kanazawa
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 |
|---|---|---|---|
| use SE(3) motion bases rather than translational bases (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D EPE: 0.093 -> 0.082 | Improved |
| use shared motion bases rather than per-Gaussian SE(3) trajectories (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D EPE: 0.083 -> 0.082 | Improved |
| use shared motion bases rather than per-Gaussian translational trajectories (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D EPE: 0.087 -> 0.082 | Improved |
| add SE(3) fitting initialization of motion bases (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D EPE: 0.111 -> 0.082 | Improved |
| add 2D track supervision (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D EPE: 0.141 -> 0.082 | Improved |
| add SE(3) motion bases on delta 10cm vs translational bases (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D delta 10cm: 69.9 -> 73.3 | Improved |
| use shared motion bases rather than per-Gaussian SE(3) on delta 5cm (ablation) | Shape of Motion · iPhone dataset (ablation, 3D tracking) | 3D delta 5cm: 43.6 -> 43.0 | Hurt |
| train with 2D Gaussian splatting plus normal supervision | Shape of Motion · iPhone dataset | 3D EPE: 0.082 -> 0.097 | Mixed |
| consolidate TAPIR+DA tracks into 4D representation | TAPIR+DA · iPhone dataset | 3D EPE: 0.114 -> 0.082 | Improved |
| compare with Dynamic Gaussian Marbles on NVIDIA | DGM · NVIDIA dataset (7 scenes, mean) | PSNR: 23.46 -> 23.37 | No clear effect |
Closest papers by meaning
- World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video (NeurIPS 2026 · similarity 0.93)
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.90)
- Reconstruct, Inpaint, Test-Time Finetune: Dynamic Novel-view Synthesis from Monocular Videos (NeurIPS 2025 · similarity 0.90)
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.89)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.89)
- OSN: Infinite Representations of Dynamic 3D Scenes from Monocular Videos (ICML 2024 · similarity 0.89)
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.
