Research map / Image generation / Image restoration
Deep Wiener Deconvolution: Wiener Meets Deep Learning for Image Deblurring
Read the paperCheck 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 |
|---|---|---|---|
| perform Wiener deconvolution on learned deep features instead of image intensity (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 27.48 -> 34.78 | Helped |
| add multi-scale encoder-decoder refinement instead of basic reconstruction (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 34.78 -> 36.90 | Helped |
| add Wiener deconvolution step (kernel guidance) to feature extractor plus refinement (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 25.70 -> 36.90 | Helped |
| add deep feature extraction before Wiener deconvolution (vs intensity only) (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 35.13 -> 36.90 | Helped |
| use learned deep features instead of image gradients only (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 28.79 -> 36.90 | Helped |
| use learned deep features instead of intensity plus gradients (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 36.52 -> 36.90 | Helped |
| increase learned features from 9 to 16 (ablation) | Deep Wiener deconvolution · Levin et al. 1% noise | PSNR (dB): 36.81 -> 36.90 | Helped |
| add multi-scale feature refinement (two scales) (ablation) | Deep Wiener deconvolution · Dataset [24], kernel 53-59, 1% noise | PSNR (dB): 27.45 -> 27.73 | Helped |
| add multi-scale feature refinement (two scales) (ablation) | Deep Wiener deconvolution · Dataset [24], kernel 13-19, 1% noise | PSNR (dB): 33.46 -> 33.57 | Helped |
| train with mix of GT and noisy kernels instead of GT kernels only (ablation) | Deep Wiener deconvolution · Dataset [18], estimated test kernels [46] | PSNR (dB): 23.74 -> 23.80 | No clear effect |
Closest papers by meaning
- Deep Non-Blind Deconvolution via Generalized Low-Rank Approximation (NeurIPS 2018 · similarity 0.89)
- Neural Blind Deconvolution Using Deep Priors (CVPR 2020 · similarity 0.88)
- Efficient Dynamic Scene Deblurring Using Spatially Variant Deconvolution Network With Optical Flow Guided Training (CVPR 2020 · similarity 0.88)
- 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 0.86)
- 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.