Research map / Image generation / Image restoration
Reflection Separation using a Pair of Unpolarized and Polarized Images
Read the paperCheck 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 |
|---|---|---|---|
| add refinement network after polarization-guided separation (ablation) | Ours · Synthetic validation (5000 pairs) | Transmission PSNR: 21.61 -> 28.23 | Helped |
| add refinement network after polarization-guided separation (ablation) | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 13.90 -> 20.92 | Helped |
| add semireflector orientation estimation (alpha, beta) instead of direct xi/zeta estimation (ablation) | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 18.84 -> 20.92 | Helped |
| add semireflector orientation estimation (alpha, beta) instead of direct xi/zeta estimation (ablation) | Ours · Synthetic validation (5000 pairs) | Transmission PSNR: 27.47 -> 28.23 | Helped |
| add gradient loss (ablation) | Ours · Synthetic validation (5000 pairs) | Transmission PSNR: 27.82 -> 28.23 | Helped |
| add gradient loss (ablation) | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 21.94 -> 20.92 | Hurt |
| feed xi and zeta into refinement network (ablation) | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 20.02 -> 20.92 | Helped |
| use our polarization-guided separation instead of ReflectNet output as refinement input (ablation) | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 18.30 -> 20.92 | Helped |
| increase input Gaussian noise to 16% (ablation) | Ours · Synthetic validation (5000 pairs) | Transmission PSNR: 28.23 -> 27.17 | Hurt |
| use pair of unpolarized+polarized images vs ReflectNet with three polarized images | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 18.50 -> 20.92 | Helped |
| test on curved parabola semireflector instead of planar (ablation) | Ours · Synthetic validation (5000 pairs) | Reflection PSNR: 20.92 -> 13.69 | Hurt |
Closest papers by meaning
- PolarFree: Polarization-based Reflection-Free Imaging (CVPR 2025 · similarity 0.88)
- Glossy Object Reconstruction with Cost-effective Polarized Acquisition (CVPR 2025 · similarity 0.86)
- RGB-to-Polarization Estimation: A New Task and Benchmark Study (NeurIPS 2025 · similarity 0.85)
- Flash-Split: 2D Reflection Removal with Flash Cues and Latent Diffusion Separation (CVPR 2025 · similarity 0.85)
- Learning to Remove Wrinkled Transparent Film with Polarized Prior (CVPR 2024 · similarity 0.85)
- NeRSP: Neural 3D Reconstruction for Reflective Objects with Sparse Polarized Images (CVPR 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.