Research map / 3D vision / Gaussian splatting
Implicit 4D Gaussian Splatting for Fast Motion with Large Inter-Frame Displacements
Seung-gyeom Kim, Areum Kim, Yongjae Yoo, Sukmin Yun
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 spatiotemporal slicing (ablation) | SPIN-4DGS · Panoptic Football | PSNR: 27.48 -> 28.96 | Improved |
| add spatiotemporal slicing (ablation) | SPIN-4DGS · Panoptic Football | train cost: 18GB -> 9GB | Improved |
| add trainable positions to hash encoder (ablation) | SPIN-4DGS network · Panoptic Basketball | PSNR: 18.64 -> 19.17 | No clear effect |
| add input position normalization (ablation) | SPIN-4DGS network · Panoptic Basketball | PSNR: 19.17 -> 29.89 | Improved |
| replace ReLU with GELU (ablation) | SPIN-4DGS network · Panoptic Basketball | PSNR: 29.89 -> 30.05 | Improved |
| enlarge hash map from 2^21 to 2^23 (ablation) | SPIN-4DGS network · Panoptic Basketball | PSNR: 30.05 -> 30.25 | Improved |
| extend early Realtime-4DGS training from 15K to 30K iterations (ablation) | SPIN-4DGS positions · Panoptic Basketball | PSNR: 29.57 -> 29.39 | Hurt |
| increase refinement from 0.5K to 2K iterations (ablation) | SPIN-4DGS · Panoptic Basketball | PSNR: 29.86 -> 30.05 | Improved |
| apply SPIN-4DGS attribute learning to D3DGS positions | D3DGS · Panoptic Softball | PSNR: 28.43 -> 30.92 | Improved |
| apply SPIN-4DGS attribute learning to Realtime-4DGS positions | Realtime-4DGS · Panoptic Basketball | PSNR: 27.89 -> 29.57 | Improved |
| apply SPIN-4DGS attribute learning to Realtime-4DGS positions | Realtime-4DGS · Panoptic Boxes | SSIM: 0.93 -> 0.92 | Hurt |
| use SPIN-4DGS instead of Realtime-4DGS | Realtime-4DGS · Neu3DV | PSNR: 32.01 -> 32.19 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- Grid4D — Grid4D: 4D Decomposed Hash Encoding for High-Fidelity Dynamic Gaussian Splatting (NeurIPS 2024) · better on 2 of 4 shared results
- MoDec-GS — MoDec-GS: Global-to-Local Motion Decomposition and Temporal Interval Adjustment for Compact Dynamic 3D Gaussian Splatting (CVPR 2025) · better on 2 of 2 shared results
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
- Orientation-anchored Hyper-Gaussian for 4D Reconstruction from Casual Videos (NeurIPS 2025 · similarity 0.91)
- Fully Explicit Dynamic Gaussian Splatting (NeurIPS 2024 · similarity 0.90)
- MotionGS: Exploring Explicit Motion Guidance for Deformable 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.89)
- Gaussian Splatting Lucas-Kanade (ICLR 2025 · similarity 0.89)
- 1000+ FPS 4D Gaussian Splatting for Dynamic Scene Rendering (NeurIPS 2025 · similarity 0.89)
- MoDec-GS: Global-to-Local Motion Decomposition and Temporal Interval Adjustment for Compact Dynamic 3D Gaussian Splatting (CVPR 2025 · 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.
