Research map / 3D vision / Depth and stereo
Estimating Spatially-Varying Lighting in Urban Scenes with Disentangled Representation
Jiajun Tang, Yongjie Zhu, Haoyu Wang, Jun Hoong Chan, Si Li, Boxin Shi
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 panoramic cosine mask Mcos (ablation) | local encoder-renderer · reconstruction | MAE: 0.034 -> 0.028 | Improved |
| add cross rendering loss Lcr (ablation) | local encoder-renderer · cross rendering | MAE: 0.048 -> 0.031 | Improved |
| add local identity loss Lid (ablation) | local encoder-renderer · cross rendering | RMSE: 0.089 -> 0.079 | Improved |
| add cross rendering loss Lcr (ablation) | local encoder-renderer · reconstruction | MAE: 0.029 -> 0.028 | No clear effect |
| add sun visibility mask Mvis (ablation) | local lighting estimator · local panorama | MAE: 0.140 -> 0.128 | Improved |
| add sun visibility mask Mvis (ablation) | local lighting estimator · local relighting | RMSE: 0.159 -> 0.145 | Improved |
| use disentangled representation instead of 5th-order SH | local lighting estimator · local panorama | MAE: 0.190 -> 0.128 | Improved |
| use disentangled representation instead of 24 SG lobes | local lighting estimator · local relighting | MAE: 0.093 -> 0.075 | Improved |
| use SH parametrization instead of disentangled representation | local lighting estimator · local panorama | RMSE: 2.394 -> 1.943 | Mixed |
| compare against SkyNet for relighting | global lighting estimator · global relighting | MAE: 0.226 -> 0.098 | Improved |
| compare against SOLID-Net for global panorama | global lighting estimator · global panorama | MAE: 0.384 -> 0.439 | Hurt |
| compare against SOLID-Net local lighting | local lighting estimator · local relighting | MAE: 0.186 -> 0.075 | Improved |
Closest papers by meaning
- Spatially-Varying Outdoor Lighting Estimation From Intrinsics (CVPR 2021 · similarity 0.91)
- Hierarchical Disentangled Representation Learning for Outdoor Illumination Estimation and Editing (ICCV 2021 · similarity 0.90)
- Spatio-Temporal Outdoor Lighting Aggregation on Image Sequences Using Transformer Networks (IJCV 2022 · similarity 0.90)
- LuxDiT: Lighting Estimation with Video Diffusion Transformer (NeurIPS 2025 · similarity 0.87)
- LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relighting (CVPR 2025 · similarity 0.86)
- Do Image Editing Models Understand Lighting? (NeurIPS 2026 · similarity 0.86)
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.
