Research map / Image generation / Image restoration
Neural Blind Deconvolution Using Deep Priors
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 |
|---|---|---|---|
| Use joint rather than alternating optimization (ablation) | SelfDeblur on the Levin dataset · Levin dataset | PSNR: 30.53 -> 33.07 | Helped |
| Use joint rather than alternating optimization (ablation) | SelfDeblur on the Levin dataset · Levin dataset | SSIM: 0.8748 -> 0.9313 | Helped |
| Add a hidden layer to the kernel generator (ablation) | SelfDeblur · Levin dataset | PSNR: 28.37 -> 33.07 | Helped |
| Add an extra hidden layer to the kernel generator (ablation) | SelfDeblur · Levin dataset | PSNR: 33.07 -> 30.92 | Hurt |
| Use an image-style asymmetric autoencoder for the kernel generator instead of an FCN (ablation) | SelfDeblur · Levin dataset | PSNR: 33.07 -> 21.51 | Hurt |
| Apply non-blind deconvolution after SelfDeblur | SelfDeblur · Levin dataset | PSNR: 33.07 -> 33.32 | Helped |
| Use SelfDeblur rather than the strongest reported baseline | Blind deconvolution · Levin dataset | PSNR: 32.99 -> 33.07 | Helped |
| Use SelfDeblur rather than the strongest reported baseline | Blind deconvolution · Levin dataset | SSIM: 0.9332 -> 0.9313 | Hurt |
| Use SelfDeblur rather than Xu & Jia | Blind deconvolution · Lai dataset average | PSNR: 20.18 -> 20.97 | Helped |
| Use SelfDeblur rather than Pan-DCP | Blind deconvolution · Lai dataset average | SSIM: 0.6656 -> 0.7524 | Helped |
Closest papers by meaning
- Non-Blind Deblurring: Handling Kernel Uncertainty With CNNs (CVPR 2018 · similarity 0.91)
- Learning Fully Convolutional Networks for Iterative Non-Blind Deconvolution (CVPR 2017 · similarity 0.91)
- Deblurring Using Analysis-Synthesis Networks Pair (CVPR 2020 · similarity 0.91)
- Cross-Scale Self-Supervised Blind Image Deblurring via Implicit Neural Representation (NeurIPS 2024 · similarity 0.89)
- DeblurDiff: Real-Word Image Deblurring with Generative Diffusion Models (NeurIPS 2025 · similarity 0.88)
- Unsupervised Blind Image Deblurring Based on Self-Enhancement (CVPR 2024 · similarity 0.88)
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.