Research map / Label-efficient and robust learning / Self-supervised and contrastive
Self-supervised Outdoor Scene Relighting
Ye Yu, Abhimitra Meka, Mohamed Elgharib, Hans-Peter Seidel, Christian Theobalt, William A. P. Smith
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 the proposed neural-rendering relighting instead of Barron and Malik | relighting pipeline · BigTime time-lapse (15 sequences) | l1 error: 0.115 -> 0.103 | Improved |
| use the proposed method instead of Barron and Malik | relighting pipeline · BigTime time-lapse (15 sequences) | MSE: 0.24 -> 0.021 | Improved |
| use the proposed method instead of Barron and Malik | relighting pipeline · BigTime time-lapse (15 sequences) | SSIM: 0.770 -> 0.760 | Hurt |
| use the proposed method instead of Yu and Smith | relighting pipeline · Relighting benchmark (original lighting 2) | SSIM: 0.780 -> 0.850 | Improved |
| use the proposed method instead of Philip et al. (multi-view, sun model) | relighting pipeline · Relighting benchmark (original lighting 6) | l1 colour error: 0.083 -> 0.073 | Improved |
| use the proposed method instead of Barron and Malik | relighting pipeline · Relighting benchmark (original lighting 1) | SSIM: 0.871 -> 0.879 | Hurt |
| add shadow channel s(p) to inverse rendering (ablation) | InverseRenderNet · qualitative MegaDepth | albedo shadow baking | Improved |
| add cycle consistency loss to direct supervision and adversarial loss (ablation) | neural renderer · MegaDepth training | stability under smooth illumination change | Improved |
| compute cross projection loss at full resolution instead of downscaling by 4 (ablation) | neural renderer · MegaDepth training | sharpness | Hurt |
| add self-reconstruction-only pre-training phase before cycle and adversarial losses (ablation) | neural renderer · MegaDepth training | convergence speed and detail | Improved |
Closest papers by meaning
- MetaGS: A Meta-Learned Gaussian-Phong Model for Out-of-Distribution 3D Scene Relighting (NeurIPS 2025 · similarity 0.89)
- Latent Intrinsics Emerge from Training to Relight (NeurIPS 2024 · similarity 0.88)
- UniRelight: Learning Joint Decomposition and Synthesis for Video Relighting (NeurIPS 2025 · similarity 0.88)
- Neural Gaffer: Relighting Any Object via Diffusion (NeurIPS 2024 · similarity 0.88)
- UniLumos: Fast and Unified Image and Video Relighting with Physics-Plausible Feedback (NeurIPS 2025 · similarity 0.87)
- ROGR: Relightable 3D Objects using Generative Relighting (NeurIPS 2025 · 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.
