Research map / 3D vision / Face and shape
PISR: Polarimetric Neural Implicit Surface Reconstruction for Textureless and Specular Objects
Guangcheng Chen, Yicheng He, Li He, Hong Zhang
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 perspective polarimetric loss instead of orthographic (with normal loss) (ablation) | PISR · Own polarimetric dataset | L1 Chamfer distance: 0.77 -> 0.54 | Improved |
| add orthographic polarimetric loss to photometric loss (ablation) | PISR · Own polarimetric dataset | L1 Chamfer distance: 1.10 -> 0.80 | Improved |
| add perspective polarimetric loss to photometric loss (ablation) | PISR · Own polarimetric dataset | L1 Chamfer distance: 1.10 -> 0.51 | Improved |
| add normal regularization with perspective polarimetric loss (ablation) | PISR · Own polarimetric dataset | L1 Chamfer distance: 0.51 -> 0.54 | Mixed |
| add normal regularization to photometric loss only (ablation) | PISR · Own polarimetric dataset | L1 Chamfer distance: 1.10 -> 1.00 | Improved |
| add normal regularization to orthographic polarimetric loss (ablation) | PISR · Own polarimetric dataset | F-score at 1 mm: 96.60 -> 96.10 | No clear effect |
| use polarization images with polarimetric loss instead of RGB-only neural SDF methods | PISR vs NeuS · Own polarimetric dataset | L1 Chamfer distance: 1.3 -> 0.5 | Improved |
| use specular-specific designs (Ref-NeuS, NeRO) instead of plain NeuS | RGB-based methods · Own polarimetric dataset | L1 Chamfer distance: 1.3 -> 1.6 and 1.7 | Hurt |
| use hash-grid neural SDF with shallow MLPs | PISR · Optimization time | time (h): 10.8 -> 0.5 vs NeuS | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 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
- Ref-NeuS — Ref-NeuS: Ambiguity-Reduced Neural Implicit Surface Learning for Multi-View Reconstruction with Reflection (ICCV 2023) · better on 2 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
- High-Fidelity Polarimetric Implicit 3D Reconstruction with View-Dependent Physical Representation (AAAI 2025 · similarity 0.87)
- A Differential Approach to Shape from Polarisation: A Level-Set Characterisation (IJCV 2018 · similarity 0.87)
- Learning Accurate 3D Shape Based on Stereo Polarimetric Imaging (CVPR 2023 · similarity 0.87)
- NeRSP: Neural 3D Reconstruction for Reflective Objects with Sparse Polarized Images (CVPR 2024 · similarity 0.87)
- Glossy Object Reconstruction with Cost-effective Polarized Acquisition (CVPR 2025 · similarity 0.86)
- GNeRP: Gaussian-guided Neural Reconstruction of Reflective Objects with Noisy Polarization Priors (ICLR 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.
