Research map / 3D vision / Gaussian splatting
UrbanGS: Efficient and Scalable Architecture for Geometrically Accurate Large-Scene Reconstruction
Changbai Li, Haodong Zhu, Hanlin Chen, Xiuping Liang, Tongfei Chen, Shuwei Shao, Linlin Yang, Huobin Tan, Baochang Zhang
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 D-Normal regularization (ablation) | UrbanGS · Modern Building (GauU-Scene) | F1: 0.463 -> 0.503 | Improved |
| add D-Normal regularization (ablation) | UrbanGS · Modern Building (GauU-Scene) | PSNR: 25.02 -> 26.44 | Improved |
| add depth consistency regularization (ablation) | UrbanGS · Modern Building (GauU-Scene) | F1: 0.453 -> 0.503 | Improved |
| add depth consistency regularization (ablation) | UrbanGS · Modern Building (GauU-Scene) | PSNR: 24.59 -> 26.44 | Improved |
| add geometry-aware confidence weighting (ablation) | UrbanGS · Modern Building (GauU-Scene) | PSNR: 26.02 -> 26.44 | Improved |
| add geometry-aware confidence weighting (ablation) | UrbanGS · Modern Building (GauU-Scene) | F1: 0.493 -> 0.503 | Improved |
| use SAGP pruning instead of LightGaussian pruning (ablation) | UrbanGS · Russian Building (GauU-Scene) | Gaussian count (M) and F1: 3.02 -> 2.45 Gaussians (M); F1 0.528 -> 0.546 | Improved |
| use proposed partitioning instead of CityGaussian partitioning (ablation) | UrbanGS · Russian Building (GauU-Scene) | F1: 0.536 -> 0.546 | Improved |
| add partitioning strategy (ST) to baseline (ablation) | UrbanGS · Russian Building (GauU-Scene) | PSNR: 22.54 -> 24.68 | Improved |
| add SAGP on top of partitioning (ablation) | UrbanGS · Russian Building (GauU-Scene) | Gaussians (M): 6.37 -> 2.45 | Improved |
| scale blocks and GPUs from 2/2 to 8/8 (ablation) | UrbanGS · Russian Building (GauU-Scene) | PSNR: 23.43 -> 24.66 | Improved |
| swap depth and normal prior estimators (Dav2, MiDaS, Dsine, GeoWizard) (ablation) | UrbanGS · Modern Building (GauU-Scene) | PSNR: 26.44 vs 26.12 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CityGaussianV2 — CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025) · better on 7 of 7 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
- CityGS-X: A Scalable Architecture for Efficient and Geometrically Accurate Large-Scale Scene Reconstruction (ICCV 2025 · similarity 0.95)
- MetroGS: Efficient and Stable Reconstruction of Geometrically Accurate High-Fidelity Large-Scale Scenes (CVPR 2026 · similarity 0.95)
- CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025 · similarity 0.94)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.91)
- EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis (CVPR 2025 · similarity 0.90)
- FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering (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.
