Research map / 3D vision / Depth and stereo
Deep 3D Capture: Geometry and Reflectance From Sparse Multi-View Images
Check 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 pairwise view encoding with max-pooled features instead of a stacked-image U-Net | Multi-view SVBRDF estimation · Synthetic test set | Roughness L2 error: 0.0359 -> 0.0275 | Helped |
| Use pairwise view encoding with max-pooled features instead of a stacked-image U-Net | Multi-view SVBRDF estimation · Synthetic test set | Diffuse L2 error: 0.0060 -> 0.0061 | Hurt |
| Add the reference image to the warped-image input (ablation) | Multi-view SVBRDF estimation · Synthetic test set | Normal L2 error: 0.0456 -> 0.0363 | Helped |
| Add depth maps to reference and warped images (ablation) | Multi-view SVBRDF estimation · Synthetic test set | Normal L2 error: 0.0363 -> 0.0321 | Helped |
| Add lighting/viewing direction maps to reference and warped images (ablation) | Multi-view SVBRDF estimation · Synthetic test set | Normal L2 error: 0.0363 -> 0.0304 | Helped |
| Use both depth maps and lighting/viewing direction maps (ablation) | Multi-view SVBRDF estimation · Synthetic test set | Roughness L2 error: 0.0306 -> 0.0275 | Helped |
| Optimize latent reflectance features rather than independently optimizing per-vertex BRDFs | SVBRDF and geometry refinement · Real captured objects | Appearance consistency | Helped |
| Add non-rigid per-view warping during refinement (ablation) | SVBRDF and geometry refinement · Real captured objects | Edge sharpness | Helped |
| Add per-vertex BRDF refinement after latent-space optimization (ablation) | SVBRDF and geometry refinement · Real captured objects | High-frequency appearance detail | Helped |
| Use six sparse views instead of a single image | Shape and SVBRDF reconstruction · Synthetic test set | Normal L2 error: 0.1075 -> 0.0226 | Helped |
Closest papers by meaning
- Neural Inverse Rendering for General Reflectance Photometric Stereo (ICML 2018 · similarity 0.91)
- 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture Objects (CVPR 2026 · similarity 0.90)
- S$^3$-NeRF: Neural Reflectance Field from Shading and Shadow under a Single Viewpoint (NeurIPS 2022 · similarity 0.89)
- RNb-NeuS: Reflectance and Normal-based Multi-View 3D Reconstruction (CVPR 2024 · similarity 0.88)
- Diffusion Reflectance Map: Single-Image Stochastic Inverse Rendering of Illumination and Reflectance (CVPR 2024 · similarity 0.88)
- UniSDF: Unifying Neural Representations for High-Fidelity 3D Reconstruction of Complex Scenes with Reflections (NeurIPS 2024 · similarity 0.88)
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.