Research map / 3D vision / Depth and stereo
ShadowNeuS: Neural SDF Reconstruction by Shadow Ray Supervision
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 ShadowNeuS instead of DeepShadow | Single-view binary-shadow reconstruction · DeepShadow dataset | Depth L1: 0.0223 -> 0.0135 | Helped |
| Use ShadowNeuS instead of DeepShadow | Single-view binary-shadow reconstruction · Our binary shadow dataset | Depth L1: 0.5020 -> 0.1870 | Helped |
| Use ShadowNeuS instead of Li et al. | RGB-based reconstruction · Our RGB dataset | Depth L1: 0.8794 -> 0.1341 | Helped |
| Use ShadowNeuS instead of Li et al. | RGB-based reconstruction · Our RGB dataset | Normal MAE: 23.61 -> 15.03 | Helped |
| Add differentiable intersection points (ablation) | ShadowNeuS · Our RGB dataset | Depth L1: 0.2569 -> 0.1341 | Helped |
| Add differentiable intersection points (ablation) | ShadowNeuS · Our RGB dataset | Normal MAE: 18.01 -> 15.03 | Helped |
| Add boundary sampling using multiple shadow rays (ablation) | ShadowNeuS · Our RGB dataset | Depth L1: 0.3552 -> 0.1341 | Helped |
| Add boundary sampling using multiple shadow rays (ablation) | ShadowNeuS · Our RGB dataset | Normal MAE: 28.44 -> 15.03 | Helped |
| Increase input images from 3 to 50 | ShadowNeuS · Chair scene | Depth L1: 0.1427 -> 0.0074 | Helped |
| Add foreground-shadow supervision to background-shadow supervision (ablation) | ShadowNeuS · Lego scene | Depth L1: 0.05827 -> 0.02955 | Helped |
| Add background-shadow supervision to foreground-shadow supervision (ablation) | ShadowNeuS · Lego scene | Depth L1: 0.13569 -> 0.02955 | Helped |
| Add foreground-shadow supervision to background-shadow supervision (ablation) | ShadowNeuS · Lego scene | Normal MAE: 29.93 -> 19.59 | Helped |
Closest papers by meaning
- S$^3$-NeRF: Neural Reflectance Field from Shading and Shadow under a Single Viewpoint (NeurIPS 2022 · similarity 0.88)
- Depth-Supervised NeRF: Fewer Views and Faster Training for Free (CVPR 2022 · similarity 0.87)
- Dense Depth Priors for Neural Radiance Fields From Sparse Input Views (CVPR 2022 · similarity 0.86)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · similarity 0.86)
- A Pre-convolved Representation for Plug-and-Play Neural Illumination Fields (AAAI 2024 · similarity 0.85)
- Gaussian Shadow Casting for Neural Characters (CVPR 2024 · similarity 0.85)
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.