Research map / 3D vision / Gaussian splatting
CoRoGS: Contextual Gaussian Splatting for Robust Large-Deviation View Synthesis
Xin Ma, Peng Lu, Yisong Chen, Chengwei Pan, Sheng Li
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 geometric update module (ablation) | CoRoGS · KITTI Right-0.5m / Left-3m | PSNR / FID: PSNR 23.25 -> 24.96; FID 83.62 -> 68.25 | Improved |
| add semantic update module (ablation) | CoRoGS · KITTI Right-0.5m / Left-3m | PSNR / FID: PSNR 23.69 -> 24.96; FID 86.93 -> 68.25 | Improved |
| use cross-modal fusion instead of concatenation (ablation) | CoRoGS · KITTI Right-0.5m | PSNR: 23.93 -> 24.96 | Improved |
| add context-aware graph expansion (ablation) | CoRoGS · KITTI Right-0.5m | PSNR: 24.03 -> 24.96 | Improved |
| add contextual smoothness loss (ablation) | CoRoGS · KITTI Left-3m | FID: 89.65 -> 68.25 | Improved |
| add normal loss (ablation) | CoRoGS · KITTI Left-3m | FID: 79.68 -> 68.25 | Improved |
| add semantic loss (ablation) | CoRoGS · KITTI Left-3m | FID: 82.96 -> 68.25 | Improved |
| use full CoRoGS instead of base model (ablation) | 3DGS base · KITTI Right-0.5m | PSNR: 21.25 -> 24.96 | Improved |
| use CoRoGS instead of DeSiRe-GS | CoRoGS · KITTI Right-0.5m | PSNR: 22.71 -> 24.96 | Improved |
| use CoRoGS instead of DC-Gaussian under large shifts | CoRoGS · KITTI+Waymo Left-5m, Up-2m, Diagonal-5m | FID: 132.52 -> 89.15 (Left-5m) | Improved |
| use CoRoGS rendering speed vs GaussianPro/SAGS | CoRoGS · KITTI Right-0.5m | FPS: SAGS 110 vs CoRoGS 49 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- DeSiRe-GS — DeSiRe-GS: 4D Street Gaussians for Static-Dynamic Decomposition and Surface Reconstruction for Urban Driving Scenes (CVPR 2025) · better on 5 of 5 shared results
- DC-Gaussian — DC-Gaussian: Improving 3D Gaussian Splatting for Reflective Dash Cam Videos (NeurIPS 2024) · better on 4 of 4 shared results
- SAGS — SAGS: Structure-Aware 3D Gaussian Splatting (ECCV 2024) · better on 2 of 2 shared results
- VEGS — VEGS: View Extrapolation of Urban Scenes in 3D Gaussian Splatting using Learned Priors (ECCV 2024) · 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
- Quantifying and Alleviating Co-Adaptation in Sparse-View 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.91)
- CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes (ICLR 2025 · similarity 0.90)
- EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis (CVPR 2025 · similarity 0.90)
- Horizon-GS: Unified 3D Gaussian Splatting for Large-Scale Aerial-to-Ground Scenes (CVPR 2025 · similarity 0.90)
- FreeSplat: Generalizable 3D Gaussian Splatting Towards Free View Synthesis of Indoor Scenes (NeurIPS 2024 · similarity 0.90)
- CoMapGS: Covisibility Map-based Gaussian Splatting for Sparse Novel View Synthesis (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.
