Research map / 3D vision / Gaussian splatting
CityGS-X: A Scalable Architecture for Efficient and Geometrically Accurate Large-Scale Scene Reconstruction
Yuanyuan Gao, Hao Li, Jiaqi Chen, Zhengyu Zou, Zhihang Zhong, Dingwen Zhang, Xiao Sun, Junwei 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 |
|---|---|---|---|
| add batch-level RGB training (ablation) | CityGS-X · Building | PSNR: 22.41 -> 23.24; time 4h21m -> 1h48m | Improved |
| add enhanced depth-prior training after batch RGB (ablation) | CityGS-X · Building | PSNR: 23.24 -> 22.46 | Hurt |
| add batch-level geometric training on top of depth prior (ablation) | CityGS-X · Building | PSNR: 22.46 -> 22.76 | Improved |
| add depth-prior training as a transition before geometric training (ablation) | CityGS-X · Building | PSNR: 22.10 -> 22.76 | Improved |
| increase batch size from 2 to 16 with more GPUs (ablation) | CityGS-X · Rubble | PSNR: 25.58 -> 26.25 | Improved |
| increase batch size from 8 to 16 (ablation) | CityGS-X · Rubble | PSNR: 26.15 -> 26.25 | No clear effect |
| replace partition-and-merge with PH2-3D parallel multi-GPU rendering | CityGS-X · MatrixCity | PSNR: 27.23 -> 27.58 | Improved |
| CityGS-X versus partition-based methods on training time | CityGS-X · Mill19 Building | training time: 03:26 -> 03:00 vs VastGS | Improved |
| add Step 3 geometric training without Step 2 depth transition (ablation) | CityGS-X · MatrixCity | geometry quality (qualitative) | Hurt |
Later papers that compared with this method
1 later accepted paper reports CityGS-X as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- City-Level 3D Surface Reconstruction with Viewpoint Orientation Partitioning and Scene Completion (ECCV 2026 · better on 3 of 4 shared results)
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.
- Momentum-GS — Momentum-GS: Momentum Gaussian Self-Distillation for High-Quality Large Scene Reconstruction (ICCV 2025) · better on 3 of 5 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 0.93)
- FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering (CVPR 2025 · similarity 0.91)
- On Scaling Up 3D Gaussian Splatting Training (ICLR 2025 · similarity 0.91)
- DOGS: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus (NeurIPS 2024 · similarity 0.91)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.90)
- Holistic Large-Scale Scene Reconstruction via Mixed Gaussian Splatting (NeurIPS 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.
