Research map / 3D vision / Gaussian splatting
SU-RGS: Relightable 3D Gaussian Splatting from Sparse Views under Unconstrained Illuminations
Qi Zhang, Chi Huang, Qian Zhang, Nan Li, Wei Feng
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 appearance embedding to varying appearance MLP (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 17.01 -> 18.20 | Improved |
| add base color input to varying appearance MLP (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 18.20 -> 19.09 | Improved |
| add position embedding input (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 19.09 -> 19.51 | Improved |
| add ray direction input (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 19.51 -> 19.95 | Improved |
| add source-to-target semantic pseudo-label re-projection (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 18.19 -> 19.11 | Improved |
| make pseudo-label re-projection dual-direction (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 19.11 -> 19.66 | Improved |
| add hierarchical semantics to pseudo-labels (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 19.66 -> 19.95 | Improved |
| use adaptive transient perception instead of constant boundary (ablation) | SU-RGS · NeRF-OSR (24 views) | PSNR: 19.50 -> 19.95 | Improved |
| use adaptive transient perception instead of cosine similarity (ablation) | SU-RGS · NeRF-OSR (24 views) | SSIM: 0.793 -> 0.876 | Improved |
| use adaptive transient perception instead of truncation (ablation) | SU-RGS · NeRF-OSR (24 views) | LPIPS: 0.340 -> 0.244 | Improved |
| SU-RGS instead of GSIR | GSIR · NeRF-OSR (24 views) | PSNR: 18.59 -> 19.95 | Improved |
| SU-RGS instead of GSIR with fewer views | GSIR · NeRF-OSR (6 views) | PSNR: 14.54 -> 16.04 | Improved |
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.
- GSIR — GS-IR: 3D Gaussian Splatting for Inverse Rendering (CVPR 2024) · better on 6 of 6 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
- MetaGS: A Meta-Learned Gaussian-Phong Model for Out-of-Distribution 3D Scene Relighting (NeurIPS 2025 · similarity 0.92)
- SVG-IR: Spatially-Varying Gaussian Splatting for Inverse Rendering (CVPR 2025 · similarity 0.91)
- MS-GS: Multi-Appearance Sparse-View 3D Gaussian Splatting in the Wild (NeurIPS 2025 · similarity 0.91)
- Generative Sparse-View Gaussian Splatting (CVPR 2025 · similarity 0.91)
- GI-GS: Global Illumination Decomposition on Gaussian Splatting for Inverse Rendering (ICLR 2025 · similarity 0.90)
- GS-IR: 3D Gaussian Splatting for Inverse Rendering (CVPR 2024 · 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.
