Research map / 3D vision / Gaussian splatting
OccluGaussian: Occlusion-Aware Gaussian Splatting for Large Scene Reconstruction and Rendering
Shiyong Liu, Xiao Tang, Zhihao Li, Yingfan He, Chongjie Ye, Jianzhuang Liu, Binxiao Huang, Shunbo Zhou, Xiaofei Wu
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 occlusion-aware camera clustering instead of grid-based division (ablation) | OccluGaussian · OccluScene3D Gallery | PSNR: 24.58 -> 25.46 | Improved |
| add occlusion-aware clustering instead of K-means on positions (ablation) | OccluGaussian · OccluScene3D Gallery | PSNR: 24.94 -> 25.46 | Improved |
| add occlusion-aware clustering instead of Metis (ablation) | OccluGaussian · OccluScene3D Gallery | PSNR: 25.02 -> 25.46 | Improved |
| add occlusion-aware clustering instead of DBSCAN (ablation) | OccluGaussian · OccluScene3D Gallery | LPIPS: 0.111 -> 0.094 | Improved |
| add NetVLAD image features to K-means clustering (ablation) | K-means (CamP) · OccluScene3D Gallery | PSNR: failed (-) | Hurt |
| add region-based rendering (culling invisible Gaussians) (ablation) | OccluGaussian · OccluScene3D Gallery | FPS: 189.52 -> 271.79 | Improved |
| add region subdivision to region-based rendering (ablation) | OccluGaussian · OccluScene3D Gallery | FPS: 271.79 -> 288.94 | Improved |
| replace VastGaussian division with occlusion-aware division (ablation) | 3DGS optimization · OccluScene3D Canteen | PSNR: 24.60 -> 25.25 | Improved |
| replace Hierarchical-GS division with occlusion-aware division (ablation) | 3DGS optimization · Zip-NeRF Berlin | PSNR: 27.26 -> 30.37 | Improved |
| increase initial cluster number from 1 to 10 (ablation) | OccluGaussian · OccluScene3D Gallery | PSNR: 25.18 -> 25.81 | Improved |
| raise initial cluster number from 10 to 15 (ablation) | OccluGaussian · OccluScene3D Gallery | PSNR: 25.81 -> 25.33 | Hurt |
| vary initial cluster number 7, 10, 15 (ablation) | OccluGaussian · Zip-NeRF NYC | PSNR: 30.90 -> 31.33 -> 31.35 | No clear effect |
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.
- VastGaussian* — VastGaussian: Vast 3D Gaussians for Large Scene Reconstruction (CVPR 2024) · better on 5 of 5 shared results
- Zip-NeRF — Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fields (ICCV 2023) · 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
- VastGaussian: Vast 3D Gaussians for Large Scene Reconstruction (CVPR 2024 · similarity 0.92)
- CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025 · similarity 0.90)
- Holistic Large-Scale Scene Reconstruction via Mixed Gaussian Splatting (NeurIPS 2025 · 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)
- Structure Consistent Gaussian Splatting with Matching Prior for Few-shot Novel View Synthesis (NeurIPS 2024 · 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.
