Research map / Image generation / Image restoration
Convergence Analysis of MAP Based Blur Kernel Estimation
Sunghyun Cho, Seungyong Lee
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 |
|---|---|---|---|
| set sparsity weight lambda_l too small (0.00001) in the energy function used as kernel metric (ablation) | Reformulated MAP energy function · Automatic blur kernel size selection | preferred kernel: 13.0, 14.4, 16.5 | Hurt |
| set lambda_l large enough and alpha small (alpha=0.1, lambda_l=0.00064) (ablation) | Reformulated MAP energy function · Levin and Sun et al. datasets | fraction of images with energy ratio below 1 (true kernel preferred over delta) | Improved |
| make lambda_l too large (ablation) | Reformulated MAP energy function · Levin and Sun et al. datasets | distinguishing true kernel from no-blur kernel | Hurt |
| use naive MAP (no edge detection, reweighting or Bayesian estimation) instead of sophisticated methods | Multi-scale naive MAP deconvolution · Levin et al. dataset | cumulative error ratio success rate | Mixed |
Closest papers by meaning
- Normalized Blind Deconvolution (ECCV 2018 · similarity 0.91)
- A Clearer Picture of Total Variation Blind Deconvolution (TPAMI 2016 · similarity 0.91)
- Blind Deblurring for Saturated Images (CVPR 2021 · similarity 0.90)
- Kernel Diffusion: An Alternate Approach to Blind Deconvolution (ECCV 2024 · similarity 0.90)
- Blind Image Deconvolution by Automatic Gradient Activation (CVPR 2016 · similarity 0.89)
- Prior and Prediction Inverse Kernel Transformer for Single Image Defocus Deblurring (AAAI 2024 · 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.
