Research map / 3D vision / 3D generation
Multi-View Inverse Rendering for Large-Scale Real-World Indoor Scenes
Zhen Li, Lingli Wang, Mofang Cheng, Cihui Pan, Jiaqi Yang
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 |
|---|---|---|---|
| Use TBL instead of SH or SG lighting | TexIR relighting · Mirror silver sphere relighting | MAE: 0.0348 -> 0.0055 | Improved |
| Add precomputed irradiance as a hybrid lighting representation (ablation) | TBL material optimization · Large-scale indoor scenes | optimization efficiency | Improved |
| Use irradiance texture (IrT) rather than neural irradiance field (NIrF) (ablation) | Hybrid lighting representation · Large-scale indoor scenes | albedo detail and artifacts | Improved |
| Add Stage I coarse albedo initialization (ablation) | TexIR material optimization · Synthetic dataset | roughness PSNR: 10.3561 -> 20.2132 | Improved |
| Add Stage II VHL sampling and semantics-based roughness propagation (ablation) | TexIR material optimization · Synthetic dataset | roughness PSNR: 7.9627 -> 20.2132 | Improved |
| Add Stage III segmentation-based fine-tuning (ablation) | TexIR material optimization · Synthetic dataset | roughness PSNR: 17.4177 -> 20.2132 | Improved |
| Use the full three-stage strategy rather than jointly optimizing albedo and roughness (ablation) | TexIR material optimization · Synthetic dataset | roughness PSNR: 7.8012 -> 20.2132 | Improved |
| Use semantics priors for roughness propagation (ablation) | TexIR · Synthetic dataset with degraded semantic masks | roughness PSNR: 20.2132 -> 13.5650 | Hurt |
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.
- PhyIR — PhyIR: Physics-Based Inverse Rendering for Panoramic Indoor Images (CVPR 2022) · 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
- Inverse Rendering for Complex Indoor Scenes: Shape, Spatially-Varying Lighting and SVBRDF From a Single Image (CVPR 2020 · similarity 0.93)
- Learning Indoor Inverse Rendering With 3D Spatially-Varying Lighting (ICCV 2021 · similarity 0.91)
- Physically-Based Editing of Indoor Scene Lighting from a Single Image (ECCV 2022 · similarity 0.90)
- MAIR: Multi-View Attention Inverse Rendering With 3D Spatially-Varying Lighting Estimation (CVPR 2023 · similarity 0.89)
- IRIS: Inverse Rendering of Indoor Scenes from Low Dynamic Range Images (CVPR 2025 · similarity 0.89)
- VMINer: Versatile Multi-view Inverse Rendering with Near- and Far-field Light Sources (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.
