Research map / 3D vision / Depth and stereo
Multi-view Surface Reconstruction Using Normal and Reflectance Cues
Robin Bruneau, Baptiste Brument, Yvain Quéau, Jean Mélou, François Lauze, Jean‐Denis Durou, Lilian Calvet
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 L2 NVR loss instead of L1 (normals only, optimal triplets) (ablation) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.252 -> 0.194 | Improved |
| use L2 NVR loss instead of L1 (normals only, optimal triplets) (ablation) | RNb-NeuS2 · LUCES-MV | Chamfer distance (mm): 0.540 -> 0.431 | Improved |
| use optimal lighting triplets instead of canonical (L1, normals only) (ablation) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.267 -> 0.252 | Improved |
| use optimal lighting triplets instead of canonical (L2, normals only) (ablation) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.196 -> 0.194 | No clear effect |
| add reflectance maps to normals under L1 loss (ablation) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.252 -> 0.241 | Improved |
| add reflectance maps to normals under L2 loss (ablation) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.194 -> 0.221 | Hurt |
| add reflectance maps to normals under L2 loss (ablation) | RNb-NeuS2 · LUCES-MV | Chamfer distance (mm): 0.431 -> 0.409 | Mixed |
| add reflectance embedding (+) to reflectance parametrisation (ablation) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.221 -> 0.209 | Improved |
| replace SDM-UniPS normals with UniMS-PS normals | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.194 -> 0.167 | Improved |
| replace SDM-UniPS normals with UniMS-PS normals | RNb-NeuS2 · LUCES-MV | Chamfer distance (mm): 0.431 -> 0.362 | Improved |
| reduce views from 20 to 5 for the proposed method (UniMS-PS) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.167 -> 0.217 | Hurt |
| use ground-truth normals instead of estimated normals (SDM) | RNb-NeuS2 · DiLiGenT-MV | Chamfer distance (mm): 0.194 -> 0.093 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- NeuS — NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 2021) · better on 3 of 3 shared results
- NeuS2 — NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction (ICCV 2023) · better on 1 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- RNb-NeuS: Reflectance and Normal-based Multi-View 3D Reconstruction (CVPR 2024 · similarity 0.92)
- Multi-View 3D Reconstruction of a Texture-Less Smooth Surface of Unknown Generic Reflectance (CVPR 2021 · similarity 0.91)
- Robust Multiview Photometric Stereo Using Planar Mesh Parameterization (TPAMI 2016 · similarity 0.91)
- Just Look at the Image: Viewpoint-Specific Surface Normal Prediction for Improved Multi-View Reconstruction (CVPR 2016 · similarity 0.90)
- PMNI: Pose-free Multi-view Normal Integration for Reflective and Textureless Surface Reconstruction (CVPR 2025 · similarity 0.89)
- Differentiable Neural Surface Refinement for Modeling Transparent Objects (CVPR 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.
