Research map / 3D vision / Gaussian splatting
CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussians
Yang Liu, Chuanchen Luo, Lue Fan, Naiyan Wang, Junran Peng, Zhaoxiang 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 coarse global Gaussian prior to block training (ablation) | CityGS baseline data selection · Rubble | PSNR: 23.98 -> 25.01 | Improved |
| add SSIM-based data assignment (Eq. 3) (ablation) | CityGS · Rubble | PSNR: 23.43 -> 25.77 | Improved |
| add inside-block pose assignment (Eq. 4) (ablation) | CityGS · Rubble | PSNR: 25.45 -> 25.77 | Improved |
| SSIM data assignment instead of 1.5x larger area with global prior (ablation) | CityGS · Rubble | #GS (M): 15.4 -> 9.7 | Improved |
| add LoD with three detail levels (ablation) | CityGS · MatrixCity (Small City) | FPS: 21.6 -> 53.7 | Improved |
| add LoD with three detail levels (ablation) | CityGS · MatrixCity (Small City) | PSNR: 27.46 -> 27.32 | Hurt |
| block-wise instead of point-wise frustum selection (ablation) | CityGS LoD · MatrixCity (Small City) | FPS: 30.3 -> 53.7 | Improved |
| use 3x3 blocks instead of 2x2 (ablation) | CityGS · Rubble | PSNR: 24.48 -> 25.41 | Improved |
| increase SSIM threshold from 0.12 to 0.14 (ablation) | CityGS · Rubble | PSNR: 25.77 -> 25.43 | Hurt |
| CityGS vs 3DGS-dagger on PSNR in Residence and Sci-Art | CityGS · Residence | PSNR: 22.00 vs Switch-NeRF 22.57 | Mixed |
| widen LoD distance intervals to [0,250],[250,500] (ablation) | CityGS LoD · MatrixCity (Small City) | PSNR / FPS: 27.32 -> 27.39 PSNR; 53.7 -> 46.4 FPS | Mixed |
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.
- GP-NeRF — GP-NeRF: Generalized Perception NeRF for Context-Aware 3D Scene Understanding (CVPR 2024) · better on 1 of 1 shared result
- MipNeRF360 — Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields (CVPR 2022) · better on 1 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
- CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025 · similarity 0.94)
- VastGaussian: Vast 3D Gaussians for Large Scene Reconstruction (CVPR 2024 · similarity 0.92)
- LightGaussian: Unbounded 3D Gaussian Compression with 15x Reduction and 200+ FPS (NeurIPS 2024 · similarity 0.89)
- LP-3DGS: Learning to Prune 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.89)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 2025 · similarity 0.89)
- MaskGaussian: Adaptive 3D Gaussian Representation from Probabilistic Masks (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.
