Research map / Image generation / Fast diffusion sampling
Stokes-Informed Diffusion for Robust Linear Polarization Estimation
Yidong Luo, Chenggong Li, Yuchao Feng, Boxin Shi, Junchao Zhang, Xin Yuan
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 Stage-II end-to-end LoRA adaptation of VAE encoder (ablation) | GenPolar · RLP test | DoLP PSNR: 28.49 -> 29.54 | Improved |
| add Stage-II end-to-end LoRA adaptation of VAE encoder (ablation) | GenPolar · RLP test | AoP MeanAE: 23.12 -> 20.15 | Improved |
| pre-adapt VAE encoder with LoRA via reconstruction only (ablation) | GenPolar · RLP test | DoLP PSNR: 28.49 -> 27.92 | Hurt |
| pre-adapt VAE encoder with LoRA via reconstruction only (ablation) | GenPolar · RLP test | AoP MeanAE: 23.12 -> 23.49 | Hurt |
| pre-adapt VAE encoder with LoRA via reconstruction only (ablation) | VAE reconstruction · RLP test | VAE DoLP mean abs error: 0.081 -> 0.035 | Improved |
| insert LoRA into VAE encoder during Stage-I multi-step training (ablation) | GenPolar Stage I · RLP test | AoP | Hurt |
| distill to one-step generator with LoRA instead of without LoRA (ablation) | GenPolar · RLP test | AoP MeanAE | Improved |
| replace end-to-end regression with diffusion sampling (ablation) | GenPolar · RLP test | DoLP/AoP | Improved |
| add observability-aware AoP loss (tau_train 0.05) (ablation) | GenPolar · RLP test | AoP | Improved |
| predict channel-wise Stokes instead of gray-averaged Stokes (ablation) | GenPolar · RLP test | polarization cues | Improved |
| predict Stokes (S1,S2) instead of direct DoLP and sin/cos 2AoP (ablation) | GenPolar · RLP test | DoLP/AoP | Improved |
| use GenPolar instead of MAE | polarization estimation · RSP test | AoP MeanAE: 33.51 -> 27.60 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Restormer — Restormer: Efficient Transformer for High-Resolution Image Restoration (CVPR 2022) · better on 9 of 9 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
- PolarScale: A Physics-Grounded Benchmark for Radiometrically Consistent RGB-to-Stokes Estimation (NeurIPS 2026 · similarity 0.92)
- RGB-to-Polarization Estimation: A New Task and Benchmark Study (NeurIPS 2025 · similarity 0.89)
- Glossy Object Reconstruction with Cost-effective Polarized Acquisition (CVPR 2025 · similarity 0.87)
- PolarFree: Polarization-based Reflection-Free Imaging (CVPR 2025 · similarity 0.87)
- NeISF: Neural Incident Stokes Field for Geometry and Material Estimation (CVPR 2024 · similarity 0.85)
- Spectral and Polarization Vision: Spectro-polarimetric Real-world Dataset (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.
