Research map / Image generation / Image restoration
Stochastic Window Transformer for Image Restoration
Jie Xiao, Xueyang Fu, Feng Wu, Zheng-Jun Zha
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 |
|---|---|---|---|
| train with stochastic window instead of fixed window (test fixed) (ablation) | Stoformer · SPA-Data deraining | PSNR: 47.80 -> 48.85 | Improved |
| add layer expectation propagation at test on top of stochastic training (ablation) | Stoformer · SPA-Data deraining | PSNR: 48.85 -> 48.97 | Improved |
| apply layer expectation propagation at test without stochastic training (ablation) | Stoformer♢♢ · SPA-Data deraining | PSNR: 47.80 -> 46.95 | Hurt |
| train with stochastic window instead of fixed window (test fixed) (ablation) | Stoformer · GoPro deblurring | PSNR: 32.80 -> 33.17 | Improved |
| add layer expectation propagation at test on top of stochastic training (ablation) | Stoformer · GoPro deblurring | PSNR: 33.17 -> 33.24 | Improved |
| apply layer expectation propagation at test without stochastic training (ablation) | Stoformer♢♢ · GoPro deblurring | PSNR: 32.80 -> 31.62 | Hurt |
| apply layer expectation propagation at test without stochastic training (ablation) | Stoformer♢♢ · HIDE deblurring | PSNR: 30.73 -> 29.94 | Hurt |
| stochastic window training and expectation propagation vs fixed window (ablation) | Stoformer · CBSD68 color denoising sigma=25 | PSNR: 31.73 -> 32.47 | Improved |
| stochastic window training and expectation propagation vs fixed window (ablation) | Stoformer · Urban100 grayscale denoising sigma=50 | PSNR: 27.90 -> 28.72 | Improved |
| stochastic window training and expectation propagation vs fixed window (ablation) | Stoformer · Set12 grayscale denoising sigma=50 | PSNR: 27.96 -> 28.46 | Improved |
| stochastic Stoformer vs Restormer | Stoformer†† · Kodak24 color denoising sigma=25 | PSNR: 33.02 -> 33.12 | No clear effect |
| stochastic Stoformer vs MPRNet | Stoformer†† · HIDE deblurring | PSNR: 30.96 -> 30.99 | No clear effect |
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
- Bayesian Window Transformer for Image Restoration (TPAMI 2025 · similarity 0.93)
- Random Shuffle Transformer for Image Restoration (ICML 2023 · similarity 0.87)
- Adaptive Window Pruning for Efficient Local Motion Deblurring (ICLR 2024 · similarity 0.86)
- Stochastic Deep Restoration Priors for Imaging Inverse Problems (ICML 2025 · similarity 0.84)
- Correlation Matching Transformation Transformers for UHD Image Restoration (AAAI 2024 · similarity 0.84)
- Plug-and-Play image restoration with Stochastic deNOising REgularization (ICML 2024 · similarity 0.84)
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.
