Research map / Image generation / Image restoration
A Comparative Study for Single Image Blind Deblurring
Check 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 |
|---|---|---|---|
| Evaluate deblurring algorithms on real-world images instead of relying only on synthetic uniform images | Single-image blind deblurring algorithms · Real dataset versus synthetic uniform dataset | Human preference ranking | Mixed |
| Use sparse-gradient or intensity priors rather than explicit edge-based methods | Single-image blind deblurring algorithms · Real dataset | Human preference | Helped |
| Use a method that produces fewer ringing and over-sharpening artifacts, even if the result remains slightly blurry | Single-image blind deblurring algorithms · Real dataset | Human preference | Helped |
| Use Pan-14's L0 intensity prior for deblurring saturated images | Pan-14 · Real saturated images | Human preference | Helped |
| Use Xu-10's kernel refinement phase to reduce noise in blur kernels | Xu-10 · Images with large motion blur and noise | Human preference | Helped |
| Apply full-reference image quality metrics to predict human preference | Deblurred images on synthetic uniform data · Synthetic uniform | Spearman's rank correlation: IFC 0.6773; PSNR 0.4135 | Helped |
| Use the motion-deblurring-specific metric trained on synthetic uniform images to predict human preference on real images | Deblurred images · Real dataset | Spearman's rank correlation: 0.4991 on Synthetic uniform; 0.1667 on Real dataset | Hurt |
| Use conventional no-reference quality metrics to predict human preference | Deblurred images · Real and synthetic datasets | Spearman's rank correlation | Hurt |
| Use uniform-blur algorithms on non-uniform blurred images | Single-image blind deblurring algorithms · Synthetic non-uniform and real datasets | Human preference | Mixed |
Closest papers by meaning
- Unsupervised Blind Image Deblurring Based on Self-Enhancement (CVPR 2024 · similarity 0.89)
- Real-World Efficient Blind Motion Deblurring via Blur Pixel Discretization (CVPR 2024 · similarity 0.88)
- Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains (CVPR 2024 · similarity 0.87)
- ID-Blau: Image Deblurring by Implicit Diffusion-based reBLurring AUgmentation (CVPR 2024 · similarity 0.87)
- BlurDM: A Blur Diffusion Model for Image Deblurring (NeurIPS 2025 · similarity 0.86)
- Cross-Scale Self-Supervised Blind Image Deblurring via Implicit Neural Representation (NeurIPS 2024 · similarity 0.86)
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.