Research map / 3D vision / Gaussian splatting
Radiometrically Consistent Gaussian Surfels for Inverse Rendering
Kyu Beom Han, Jaeyoon Kim, Woo Jae Kim, Jinhwan Seo, Sung-eui Yoon
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 radiometric consistency loss (ablation) | RadioGS · Indirect-illumination dataset | indirect PSNR: 30.0954 -> 32.8832 | Improved |
| add radiometric consistency loss when training views reduced to 25% (ablation) | RadioGS · Indirect-illumination dataset (25% views) | indirect PSNR: 27.89 -> 32.71 | Improved |
| detach gradient through ray-traced surfels (ablation) | RadioGS · TensoIR | NVS PSNR / relighting PSNR: 37.252 -> 37.858 NVS | Improved |
| use differentiable ray-traced indirect radiance instead of frozen precomputed (ablation) | RadioGS · TensoIR | NVS PSNR: 35.547 -> 37.858 | Improved |
| add indirect radiance from surfels instead of split-sum (ablation) | RadioGS · TensoIR | albedo PSNR: 27.471 -> 31.048 | Improved |
| increase number of supervised surfels N_g from 1024 to 8192 (ablation) | RadioGS · TensoIR | relight PSNR: 32.0569 -> 32.1294 | Improved |
| increase secondary ray samples N_s from 64 to 128 (ablation) | RadioGS · TensoIR | NVS/relight PSNR | No clear effect |
| use finetuning-based relighting instead of ray-tracing relighting | RadioGS · TensoIR | relight PSNR and render time: 32.09 -> 31.41 PSNR; 38.64 -> 5.902 ms | Mixed |
| add physically-based radiometric loss during initialization (ablation) | RadioGS · TensoIR | PSNR | Improved |
| RadioGS vs best baseline | RadioGS · Synthetic4Relight | roughness MSE: 0.008 -> 0.011 | Hurt |
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.
- IRGS — IRGS: Inter-Reflective Gaussian Splatting with 2D Gaussian Ray Tracing (CVPR 2025) · better on 8 of 8 shared results
- SVG-IR — SVG-IR: Spatially-Varying Gaussian Splatting for Inverse Rendering (CVPR 2025) · better on 7 of 7 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
- MaterialRefGS: Reflective Gaussian Splatting with Multi-view Consistent Material Inference (NeurIPS 2025 · similarity 0.89)
- SVG-IR: Spatially-Varying Gaussian Splatting for Inverse Rendering (CVPR 2025 · similarity 0.89)
- GS-IR: 3D Gaussian Splatting for Inverse Rendering (CVPR 2024 · similarity 0.89)
- GI-GS: Global Illumination Decomposition on Gaussian Splatting for Inverse Rendering (ICLR 2025 · similarity 0.89)
- Volumetrically Consistent 3D Gaussian Rasterization (CVPR 2025 · similarity 0.88)
- Structure Consistent Gaussian Splatting with Matching Prior for Few-shot Novel View Synthesis (NeurIPS 2024 · similarity 0.88)
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.
