Research map / Image generation / Image restoration
Blind image deblurring with noise-robust kernel estimation
Chanseok Lee, Jeongsol Kim, Seungmin Lee, Jaehwang Jung, Yunje Cho, Taejoong Kim, Taeyong Jo, Myungjun Lee, Mooseok Jang
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 |
|---|---|---|---|
| replace prior blind deblurring methods with noise-robust kernel estimation plus DIP | blind deblurring · AFHQ-dog (sigma=20) | PSNR: 19.94 -> 22.90 | Improved |
| replace best baseline (Pure-let) with proposed method on strong noise | blind deblurring · Sun et al. dataset (sigma=20) | PSNR: 17.31 -> 24.83 | Improved |
| increase noise from sigma=10 to sigma=20 | Proposed method · AFHQ-dog | PSNR: 24.23 -> 22.90 | Hurt |
| increase noise from sigma=10 to sigma=20 | WDIP · Sun et al. dataset | PSNR: 17.60 -> 12.89 | Hurt |
| use multiple noise suppression factors k_i (5^2, 15^2, 25^2) instead of one precisely known k (ablation) | kernel estimation · AFHQ simulated blur | deblurring quality | No clear effect |
| set noise suppression factor k too small relative to noise (0.1 sigma^2) (ablation) | kernel estimation function · kernel estimation | kernel quality | Hurt |
| set noise suppression factor k too large (10 or 100 sigma^2) (ablation) | kernel estimation function · kernel estimation | kernel quality | Hurt |
Closest papers by meaning
- Robust Kernel Estimation With Outliers Handling for Image Deblurring (CVPR 2016 · similarity 0.92)
- Blind Noisy Image Deblurring Using Residual Guidance Strategy (ICCV 2025 · similarity 0.92)
- Blind Deblurring for Saturated Images (CVPR 2021 · similarity 0.92)
- Uncertainty-Aware Unsupervised Image Deblurring With Deep Residual Prior (CVPR 2023 · similarity 0.91)
- Non-Blind Deblurring: Handling Kernel Uncertainty With CNNs (CVPR 2018 · similarity 0.91)
- Blind Image Deblurring With Outlier Handling (ICCV 2017 · similarity 0.91)
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.
