Research map / 3D vision / Gaussian splatting
RetimeGS: Continuous-Time Reconstruction of 4D Gaussian Splatting
Xuezhen Wang, Li Ma, Yulin Shen, Zeyu Wang, Pedro V. Sander
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 bidirectional optical-flow supervision (ablation) | RetimeGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 27.24 -> 30.08 | Improved |
| add triple-rendering supervision (ablation) | RetimeGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 27.16 -> 30.08 | Improved |
| add dynamic stretching of static primitives (ablation) | RetimeGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 28.81 -> 30.08 | Improved |
| use Catmull-Rom spline trajectory instead of linear (ablation) | RetimeGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 28.50 -> 30.08 | Improved |
| add flow-aware initialization (ablation) | RetimeGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 29.69 -> 30.08 | Improved |
| add flow-aware initialization (ablation) | RetimeGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | LPIPS: 0.0227 -> 0.0225 | No clear effect |
| add optical flow supervision to 4D primitives (GaussianFlow vs STGS) | STGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 25.34 -> 25.91 | No clear effect |
| use RetimeGS instead of Deform-GS | 4DGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | PSNR: 28.45 -> 30.08 | Improved |
| use RetimeGS instead of 2D video-interpolation lifting | 4DGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | LPIPS: 0.0267 -> 0.0225 | Improved |
| use RetimeGS instead of STGS | 4DGS · Stage-Capture Dataset (held-out intermediate frames, foreground) | SSIM: 0.825 -> 0.904 | Improved |
Closest papers by meaning
- 1000+ FPS 4D Gaussian Splatting for Dynamic Scene Rendering (NeurIPS 2025 · similarity 0.93)
- Temporal Smoothness-Aware Rate-Distortion Optimized 4D Gaussian Splatting (NeurIPS 2025 · similarity 0.93)
- Fully Explicit Dynamic Gaussian Splatting (NeurIPS 2024 · similarity 0.92)
- Gaussian-Flow: 4D Reconstruction with Dynamic 3D Gaussian Particle (CVPR 2024 · similarity 0.91)
- Instant4D: 4D Gaussian Splatting in Minutes (NeurIPS 2025 · similarity 0.90)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 2025 · similarity 0.90)
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.
