Research map / 3D vision / 3D generation
Objects As Cameras: Estimating High-Frequency Illumination From Shadows
Tristan Swedish, Connor Henley, Ramesh Raskar
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 L2 regularization (ablation) | linear shadow inversion · Multi-illumination dots (unknown albedo, synthetic) | Envmap RMSE: 0.417 -> 0.075 | Improved |
| add L2 regularization (ablation) | linear shadow inversion · Multi-illumination dots (unknown albedo, synthetic) | Envmap SSIM: 0.43 -> 0.62 | Improved |
| add spatial smoothness on top of L2 (ablation) | linear shadow inversion · Multi-illumination dots (unknown albedo, synthetic) | Envmap SSIM: 0.62 -> 0.63 | No clear effect |
| add L2 regularization (ablation) | linear shadow inversion · Known albedo garden HDRI (single iteration) | Envmap RMSE: 0.492 -> 0.078 | Improved |
| add spatial smoothness on top of L2 (ablation) | linear shadow inversion · Known albedo garden HDRI (single iteration) | Envmap RMSE: 0.078 -> 0.063 | Improved |
| add spatial smoothness on top of L2 (ablation) | linear shadow inversion · Known albedo garden HDRI (single iteration) | Envmap SSIM: 0.77 -> 0.83 | Improved |
| add albedo regularizer of [36] (ablation) | +L2+Smooth · Multi-illumination dots (unknown albedo, synthetic) | Albedo RMSE: 0.219 -> 0.218 | No clear effect |
| add L2 regularization (ablation) | linear shadow inversion · Multi-illumination dots (unknown albedo, synthetic) | Albedo RMSE: 0.286 -> 0.219 | Improved |
| use linear solve instead of Adam gradient descent (early stop) | +L2+Smooth · Multi-illumination dots (unknown albedo, synthetic) | Envmap RMSE: 0.531 -> 0.075 | Improved |
| use linear solve instead of full-convergence gradient descent | +L2+Smooth · Multi-illumination dots (unknown albedo, synthetic) | Compute time (s): 1434 -> 327 | Improved |
| use linear solve instead of gradient descent for albedo estimation | +L2+Smooth · Multi-illumination dots (unknown albedo, synthetic) | Albedo RMSE: 0.172 -> 0.219 | Hurt |
| use linear solve instead of full gradient descent (known albedo) | +L2+Smooth · Known albedo garden HDRI (single iteration) | Envmap RMSE: 0.120 -> 0.063 | Improved |
Closest papers by meaning
- Eclipse: Disambiguating Illumination and Materials using Unintended Shadows (CVPR 2024 · similarity 0.89)
- Object-based Illumination Estimation with Rendering-aware Neural Networks (ECCV 2020 · similarity 0.88)
- Generative Multiview Relighting for 3D Reconstruction under Extreme Illumination Variation (CVPR 2025 · similarity 0.87)
- ReCap: Better Gaussian Relighting with Cross-Environment Captures (CVPR 2025 · similarity 0.87)
- Factored-NeuS: Reconstructing Surfaces, Illumination, and Materials of Possibly Glossy Objects (CVPR 2025 · similarity 0.87)
- Recasting Regional Lighting for Shadow Removal (AAAI 2024 · similarity 0.87)
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.
