Research map / 3D vision / Gaussian splatting
City-Level 3D Surface Reconstruction with Viewpoint Orientation Partitioning and Scene Completion
Liang Han, Wenyuan Zhang, Junsheng Zhou, Yushen Liu, Zhizhong Han
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 45 degree angular threshold instead of 15 degrees for view clustering (ablation) | VOP-GS · GauU-Scene | F1: 0.534 -> 0.560 | Improved |
| use 45 degree angular threshold instead of 60 degrees (ablation) | VOP-GS · GauU-Scene | F1: 0.532 -> 0.560 | Improved |
| add initial point cloud completion (tau=0.7) (ablation) | VOP-GS · GauU-Scene | F1: 0.546 -> 0.560 | Improved |
| add initial point cloud completion (tau=0.7) (ablation) | VOP-GS · MatrixCity | F1: 0.772 -> 0.803 | Improved |
| raise entropy threshold from 0.6 to 0.85 (ablation) | VOP-GS · GauU-Scene | F1: 0.551 -> 0.563 | Improved |
| raise entropy threshold from 0.6 to 0.85 (ablation) | VOP-GS · MatrixCity | F1: 0.797 -> 0.806 | Improved |
| raise entropy threshold to 0.9 (ablation) | VOP-GS · MatrixCity | F1: 0.806 -> OOM | Hurt |
| initialize with completed point cloud instead of sparse (ablation) | 2DGS · GauU-Scene (6 scenes) | Mean F1: .324 -> .336 | Improved |
| initialize with completed point cloud instead of sparse (ablation) | PGSR · GauU-Scene (6 scenes) | Mean F1: .333 -> .382 | Improved |
| initialize with completed point cloud instead of sparse (ablation) | CityGaussian-V2 · GauU-Scene (6 scenes) | Mean F1: .506 -> .514 | Improved |
| initialize with completed point cloud instead of sparse (ablation) | CityGS-X · GauU-Scene (6 scenes) | Mean F1: .527 -> .533 | Improved |
| use viewpoint-orientation grouping vs CityGS-X | VOP-GS · MatrixCity | F1: 0.581 -> 0.803 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- CityGaussian-V2 — CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025) · better on 4 of 4 shared results
- CityGS-X — CityGS-X: A Scalable Architecture for Efficient and Geometrically Accurate Large-Scale Scene Reconstruction (ICCV 2025) · better on 3 of 4 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
- CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025 · similarity 0.90)
- GigaGS: 3D Gaussian Based Planar Representation for Large-Scene Surface Reconstruction (AAAI 2025 · similarity 0.90)
- FatesGS: Fast and Accurate Sparse-View Surface Reconstruction Using Gaussian Splatting with Depth-Feature Consistency (AAAI 2025 · similarity 0.89)
- Graph-Guided Scene Reconstruction from Images with 3D Gaussian Splatting (ICLR 2025 · similarity 0.88)
- VA-GS: Enhancing the Geometric Representation of Gaussian Splatting via View Alignment (NeurIPS 2025 · similarity 0.88)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.87)
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.
