Research map / 3D vision / NeRF and radiance fields
LightCity: An Urban Dataset for Outdoor Inverse Rendering and Reconstruction under Multi-illumination Conditions
Jingjing Wang, Qirui Hu, Chong Bao, Yuke Zhu, Hujun Bao, Zhaopeng Cui, Guofeng Zhang
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 LightCity to Hypersim fine-tuning (ablation) | DMP · LightCity-Outdoor albedo | si-PSNR: 17.61 -> 22.80 | Improved |
| add LightCity to Hypersim fine-tuning (ablation) | DMP · LightCity-Outdoor shading | si-PSNR: 18.38 -> 26.58 | Improved |
| add LightCity to Hypersim fine-tuning (ablation) | DPF · LightCity-Outdoor albedo | si-PSNR: 16.20 -> 18.94 | Improved |
| add LightCity to Hypersim fine-tuning (ablation) | DPF · EDEN-Outdoor OOD shading | si-PSNR: 10.56 -> 14.16 | Improved |
| add LightCity to Hypersim fine-tuning (ablation) | DMP · EDEN-Outdoor OOD shading | si-PSNR: 8.09 -> 11.54 | Improved |
| add LightCity to Hypersim fine-tuning (ablation) | DMP · EDEN-Outdoor OOD albedo | si-PSNR: 13.12 -> 15.10 | Improved |
| add LightCity to Hypersim fine-tuning (ablation) | DMP · Hypersim-Indoor albedo | si-PSNR: 16.48 -> 16.56 | No clear effect |
| add LightCity to Hypersim fine-tuning (ablation) | DMP · IIW WHDR | WHDR (%): 19.08 -> 20.37 | Hurt |
| add LightCity to Hypersim fine-tuning (ablation) | DPF · IIW WHDR | WHDR (%): 43.14 -> 38.502 | Improved |
| fine-tune DMP with LightCity instead of Hypersim only for albedo consistency (ablation) | DMP · Real-world albedo consistency across lighting | average variance: 0.042 -> 0.015 | Improved |
| train Gaussian-wild on multi-illumination instead of single-illumination data (ablation) | Gaussian-wild · LightCity F2 NVS | PSNR: 28.74 -> 25.72 | Hurt |
| evaluate on unseen single-illumination test set instead of multi-illumination test | Gaussian-wild · LightCity F2 NVS | PSNR: 25.72 -> 21.54 | Hurt |
Closest papers by meaning
- Neural Fields Meet Explicit Geometric Representations for Inverse Rendering of Urban Scenes (CVPR 2023 · similarity 0.89)
- MatrixCity: A Large-scale City Dataset for City-scale Neural Rendering and Beyond (ICCV 2023 · similarity 0.88)
- SkyLume: A Large-Scale Multi-Illumination Aerial Benchmark for Urban Scene Reconstruction and Beyond (ECCV 2026 · similarity 0.87)
- InvRGB+L: Inverse Rendering of Complex Scenes with Unified Color and LiDAR Reflectance Modeling (ICCV 2025 · similarity 0.86)
- Multi-Level Neural Scene Graphs for Dynamic Urban Environments (CVPR 2024 · similarity 0.85)
- Latent Intrinsics Emerge from Training to Relight (NeurIPS 2024 · similarity 0.84)
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.
