Research map / 3D vision / NeRF and radiance fields
Deep Reflectance Volumes: Relightable Reconstructions from Multi-View Photometric Images
Sai Bi, Zexiang Xu, Kalyan Sunkavalli, Miloš Hašan, Yannick Hold-Geoffroy, David Kriegman, Ravi Ramamoorthi
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 |
|---|---|---|---|
| reduce training images on House scene below 100 (ablation) | Deep Reflectance Volumes · House scene | SSIM/PSNR | Hurt |
| use 385 instead of 200 training images on House scene (ablation) | Deep Reflectance Volumes · House scene | SSIM/PSNR: 385 images vs 200 images | No clear effect |
| add deep volume prior (decoder network) instead of direct voxel optimization (ablation) | Deep Reflectance Volumes · optimization convergence | reconstruction loss convergence speed | Improved |
| train 450000 iterations instead of 200000 (ablation) | Deep Reflectance Volumes · training schedule | reconstruction quality: 200000 iterations for 1 day vs 450000 iterations for about two days | No clear effect |
| replace mesh-based per-vertex BRDF method (Nam et al.) with volumetric reflectance | Deep Reflectance Volumes · relighting and view synthesis on real scenes | visual quality | Improved |
| replace DeepVoxels with Deep Reflectance Volumes | Deep Reflectance Volumes · novel view synthesis, collocated light, 512x512 | PSNR/SSIM | Improved |
Closest papers by meaning
- Neural Multi-View Self-Calibrated Photometric Stereo without Photometric Stereo Cues (ICCV 2025 · similarity 0.90)
- Deep 3D Capture: Geometry and Reflectance From Sparse Multi-View Images (CVPR 2020 · similarity 0.90)
- Neural Inverse Rendering for General Reflectance Photometric Stereo (ICML 2018 · similarity 0.90)
- RNb-NeuS: Reflectance and Normal-based Multi-View 3D Reconstruction (CVPR 2024 · similarity 0.88)
- Differentiable Neural Surface Refinement for Modeling Transparent Objects (CVPR 2024 · similarity 0.87)
- Generative Multiview Relighting for 3D Reconstruction under Extreme Illumination Variation (CVPR 2025 · 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.
