Research map / Image generation / Image restoration
Learning Data Terms for Non-blind Deblurring
Jiangxin Dong, Jinshan Pan, Deqing Sun, Zhixun Su, Ming-Hsuan Yang
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 |
|---|---|---|---|
| learn the data term shrinkage function instead of fixed l2 norm (ablation) | TV-prior deblurring, impulse noise example · Impulse noise example image (Nf=0) | PSNR: 19.26 -> 29.99 | Improved |
| learn the data term shrinkage function instead of fixed l1 norm (ablation) | TV-prior deblurring, impulse noise example · Impulse noise example image (Nf=0) | PSNR: 27.36 -> 29.99 | Improved |
| use stage-dependent instead of stage-independent data term parameters (ablation) | Ours · Impulse noise example image | visual quality | Improved |
| combine the learned data term with the EPLL prior | EPLL · Images with outliers | visual quality | Improved |
| learned data term versus hand-crafted Cho model on impulse noise (1%) | Cho · Impulse noise 1% | PSNR: 27.73 -> 29.75 | Improved |
| learned data term versus EPLL on low Gaussian noise (1%) | EPLL · Gaussian noise 1% | PSNR: 27.64 -> 26.47 | Hurt |
| learned data term versus EPLL at high Gaussian noise (5%) | EPLL · Gaussian noise 5% | PSNR: 23.56 -> 23.73 | Improved |
| learned data term versus Cho on mixed noise 1% | Cho · Mixed noise 1% | PSNR: 27.15 -> 26.23 | Hurt |
| learned data term versus Cho on saturated pixels | Cho · Saturated pixels | PSNR: 28.17 -> 29.55 | Improved |
Closest papers by meaning
- Learning Discriminative Data Fitting Functions for Blind Image Deblurring (ICCV 2017 · similarity 0.91)
- Noise-Blind Image Deblurring (CVPR 2017 · similarity 0.90)
- Learning a Discriminative Prior for Blind Image Deblurring (CVPR 2018 · similarity 0.89)
- Unsupervised Blind Image Deblurring Based on Self-Enhancement (CVPR 2024 · similarity 0.89)
- Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains (CVPR 2024 · similarity 0.87)
- DeblurDiff: Real-Word Image Deblurring with Generative Diffusion Models (NeurIPS 2025 · similarity 0.87)
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.
