Research map / 3D vision / Face and shape
Inverse Rendering for Complex Indoor Scenes: Shape, Spatially-Varying Lighting and SVBRDF From a Single Image
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Represent lighting with spherical Gaussians instead of spherical harmonics | Indoor-scene lighting approximation · Our test set | Lighting approximation log L2: 4.43 -> 1.56 | Helped |
| Represent lighting with spherical Gaussians instead of spherical harmonics | Indoor-scene lighting approximation · Our test set | Rendered image L2: 8.6\times 10^{-3} -> 7.6\times 10^{-3} | Helped |
| Add the second cascade stage (ablation) | Inverse rendering network · IIW | WHDR: 23.29 -> 21.99 | Helped |
| Add the second cascade stage (ablation) | Inverse rendering network fine-tuned on IIW · IIW | WHDR: 16.83 -> 15.93 | Helped |
| Fine-tune the model on IIW | Ours (cascade 1) · IIW | WHDR: 21.99 -> 15.93 | Helped |
| Add the second cascade stage (ablation) | Inverse rendering network · NYU | Normal mean angular error: 27.08 -> 26.33 | Helped |
| Add the second cascade stage (ablation) | Inverse rendering network · NYU | Depth (Inv.): 0.217 -> 0.206 | Helped |
| Jointly train material, geometry, and lighting rather than training them independently (ablation) | Inverse rendering network · Our test set | Lighting log L2 error: 2.54 -> 2.50 | Helped |
| Jointly train material, geometry, and lighting rather than training them independently (ablation) | Inverse rendering network · Our test set | Normal error (10^{-2}): 4.91 -> 4.51 | Helped |
| Add predicted material and geometry as inputs to the lighting network (ablation) | Lighting prediction network · Our test set | Lighting log L2 error: 2.83 -> 2.54 | Helped |
| Add ground-truth spherical Gaussian parameter supervision (ablation) | Lighting prediction network · Our test set | Rendered image error (10^{-2}): 1.56 -> 1.06 | Helped |
| Add bilateral solver refinement (ablation) | Cascade 1 material and geometry predictions · Our test set | Albedo error (10^{-2}): 1.18 -> 1.16 | Helped |
Closest papers by meaning
- PhyIR: Physics-Based Inverse Rendering for Panoramic Indoor Images (CVPR 2022 · similarity 0.93)
- IRIS: Inverse Rendering of Indoor Scenes from Low Dynamic Range Images (CVPR 2025 · similarity 0.90)
- Channel-wise Noise Scheduled Diffusion for Inverse Rendering in Indoor Scenes (CVPR 2025 · similarity 0.89)
- Factored-NeuS: Reconstructing Surfaces, Illumination, and Materials of Possibly Glossy Objects (CVPR 2025 · similarity 0.88)
- VMINer: Versatile Multi-view Inverse Rendering with Near- and Far-field Light Sources (CVPR 2024 · similarity 0.88)
- Eclipse: Disambiguating Illumination and Materials using Unintended Shadows (CVPR 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.