Research map / Image generation / Image restoration
Learning to Deblur
Christian J. Schuler, Michael Hirsch, Stefan Harmeling B, Bernhard Schölkopf
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 |
|---|---|---|---|
| add more stacked stages (iterations) to the network (ablation) | Learned blind deconvolution NN · kernel estimation | kernel estimation performance | Improved |
| add multi-scale procedure over single-scale network (ablation) | Learned blind deconvolution NN · Levin et al. benchmark | kernel estimation performance | Improved |
| train on a single image category (valley, blackboard) instead of all ImageNet | Learned blind deconvolution NN · valley and blackboard test images | deblurring quality (qualitative) | Improved |
| train with 5% Gaussian noise instead of 1% | Learned blind deconvolution NN · image with 5% Gaussian noise | PSNR | Improved |
| increase blur kernel size beyond 25x25 | Learned blind deconvolution NN · Levin et al. benchmark | deblurring accuracy | Hurt |
| increase size of input image crop | Learned blind deconvolution NN · 52 largest images of Sun et al. | MSE | Improved |
Closest papers by meaning
- Neural Blind Deconvolution Using Deep Priors (CVPR 2020 · similarity 0.90)
- Learning a Discriminative Prior for Blind Image Deblurring (CVPR 2018 · similarity 0.90)
- Learning Fully Convolutional Networks for Iterative Non-Blind Deconvolution (CVPR 2017 · similarity 0.89)
- Douglas-Rachford Networks: Learning Both the Image Prior and Data Fidelity Terms for Blind Image Deconvolution (CVPR 2019 · similarity 0.89)
- Deep Wiener Deconvolution: Wiener Meets Deep Learning for Image Deblurring (NeurIPS 2020 · similarity 0.89)
- DWDN: Deep Wiener Deconvolution Network for Non-Blind Image Deblurring (TPAMI 2021 · similarity 0.89)
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.
