Research map / 3D vision / 3D generation
PolarAnything: Diffusion-based Polarimetric Image Synthesis
Kailong Zhang, Youwei Lyu, Heng Guo, Si Li, Zhanyu Ma, Boxin Shi
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 |
|---|---|---|---|
| encode AoLP as (cos2Phi, sin2Phi) instead of raw AoLP (ablation) | PolarAnything · own real-world test set | MAngE: 29.46 -> 25.33 | Improved |
| encode AoLP as (cos2Phi, sin2Phi) instead of raw AoLP (ablation) | PolarAnything · own real-world test set | PSNR: 40.57 -> 41.74 | Improved |
| encode AoLP as (cos2Phi, sin2Phi) instead of raw AoLP (ablation) | PolarAnything · own real-world test set | MAbsE: 0.1100 -> 0.1075 | Improved |
| generate AoLP and DoLP instead of four polarization images (ablation) | PolarAnything · own real-world test set | MAngE: 45.67 -> 29.46 | Improved |
| generate AoLP and DoLP instead of four polarization images (ablation) | PolarAnything · own real-world test set | PSNR: 23.23 -> 40.57 | Improved |
| expand DeepSfP training with PolarStanford-ORB (300 synthesized images) | DeepSfP · PN test set | mean angular error: 30.69 -> 22.93 | Improved |
| expand DeepSfP training with PolarStanford-ORB (300 synthesized images) | DeepSfP · DP test set | mean angular error: 19.38 -> 19.20 | Improved |
| expand DeepSfP training with PolarStanford-ORB (300 synthesized images) | DeepSfP · DP test set | median angular error: 15.21 -> 15.71 | Hurt |
| expand DeepSfP training with PolarStanford-ORB (300 synthesized images) | DeepSfP · DP+PN test set | mean angular error: 22.21 -> 20.13 | Improved |
| use PolarAnything-generated PSO instead of Mitsuba-rendered MSO for DeepSfP expansion | DeepSfP · PN test set | mean angular error: 24.22 -> 22.93 | Improved |
| use PolarAnything-generated PSO instead of Mitsuba-rendered MSO for DeepSfP expansion | DeepSfP · DP test set | mean angular error: 19.15 -> 19.20 | No clear effect |
| use diffusion backbone instead of Restormer (ablation) | PolarAnything · polarization synthesis | AoLP and DoLP quality | Improved |
Closest papers by meaning
- Stokes-Informed Diffusion for Robust Linear Polarization Estimation (ECCV 2026 · similarity 0.88)
- Glossy Object Reconstruction with Cost-effective Polarized Acquisition (CVPR 2025 · similarity 0.87)
- PolarFree: Polarization-based Reflection-Free Imaging (CVPR 2025 · similarity 0.87)
- PolarScale: A Physics-Grounded Benchmark for Radiometrically Consistent RGB-to-Stokes Estimation (NeurIPS 2026 · similarity 0.86)
- RGB-to-Polarization Estimation: A New Task and Benchmark Study (NeurIPS 2025 · similarity 0.85)
- High-Fidelity Polarimetric Implicit 3D Reconstruction with View-Dependent Physical Representation (AAAI 2025 · 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.
