Research map / 3D vision / 3D generation
Street Gaussians without 3D Object Tracker
Ruida Zhang, Chengxi Li, Chenyangguang Zhang, Xingyu Liu, Haili Yuan, Yanyan Li, Xiangyang Ji, Gim Hee Lee
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 2D-tracker-based motion modeling instead of no tracking (ablation) | Ours · Waymo-NOTR NVS | PSNR: 25.25 -> 28.85 | Improved |
| add motion loss (ablation) | Ours · Waymo-NOTR NVS | PSNR: 28.39 -> 28.85 | Improved |
| add isolation (ablation) | Ours · Waymo-NOTR NVS | PSNR: 28.58 -> 28.85 | Improved |
| replace 3D tracker with 2D tracker (ablation) | Ours · Waymo-NOTR NVS | PSNR: 28.02 -> 28.85 | Improved |
| add motion loss to 3D tracker variant (ablation) | 3D tracker variant · Waymo-NOTR NVS | PSNR: 27.54 -> 28.02 | Improved |
| use GT pose instead of 2D tracker (ablation) | Ours · Waymo-NOTR NVS | PSNR: 28.85 -> 28.98 | Mixed |
| use UNINEXT as tracker (ablation) | Ours · Waymo-NOTR NVS | PSNR: 28.76 vs 28.85 | No clear effect |
| use HexPlane motion instead of rigid transform (ablation) | Ours · Waymo-NOTR NVS | PSNR: 27.69 -> 28.85 | Improved |
| use VoxelNext instead of CasTrack 3D tracker (ablation) | Street Gaussians · Waymo-NOTR NVS | PSNR: 25.61 -> 26.98 | Improved |
| our method vs Street Gaussians on KITTI-75% | Ours · KITTI-75% | PSNR: 25.79 -> 25.49 | Hurt |
| our method vs Street Gaussians | Ours · Waymo-NOTR NVS | PSNR: 26.98 -> 28.85 | Improved |
| runtime cost vs Street Gaussians | Ours · inference | FPS: 160 -> 100 | Hurt |
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.
- EmerNeRF — EmerNeRF: Emergent Spatial-Temporal Scene Decomposition via Self-Supervision (ICLR 2024) · better on 1 of 1 shared result
- Street Gaussians — Street Gaussians: Modeling Dynamic Urban Scenes with Gaussian Splatting (ECCV 2024) · better on 0 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
- Monocular Quasi-Dense 3D Object Tracking (TPAMI 2022 · similarity 0.89)
- Street Gaussians: Modeling Dynamic Urban Scenes with Gaussian Splatting (ECCV 2024 · similarity 0.88)
- Street Gaussians: Modeling Dynamic Urban Scenes With Gaussian Primitives (TPAMI 2025 · similarity 0.88)
- RegTrack: Simplicity Beneath Complexity in Robust Multi-Modal 3D Multi-Object Tracking (TPAMI 2026 · similarity 0.87)
- 3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline (ICLR 2025 · similarity 0.87)
- S2-Track: A Simple yet Strong Approach for End-to-End 3D Multi-Object Tracking (ICML 2025 · similarity 0.86)
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.
