Research map / Image generation / Image restoration
Variational-EM-Based Deep Learning for Noise-Blind Image Deblurring
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 the noise-blind VEM-Net instead of the strongest noise-blind baseline | image deblurring · Set12, σ=7.65 | PSNR (dB): 27.89 -> 28.47 | Helped |
| use the noise-blind VEM-Net instead of the strongest noise-blind baseline | image deblurring · Sun et al.'s, σ=7.65 | PSNR (dB): 28.97 -> 29.41 | Helped |
| use the noise-blind VEM-Net instead of the strongest noise-blind baseline | image deblurring · Levin et al.'s, σ=7.65 | PSNR (dB): 29.79 -> 30.50 | Helped |
| add M-step parameter prediction instead of using fixed β (ablation) | VEM-Net · Set12 | PSNR (dB): 31.14 -> 31.93; 29.38 -> 29.78; 28.15 -> 28.47; 27.29 -> 27.52; 26.52 -> 26.77 | Helped |
| add CNN-based prediction of λ instead of using tuned fixed λ (ablation) | VEM-Net · Set12 | PSNR (dB): 31.36 -> 31.93; 29.47 -> 29.78; 28.24 -> 28.47; 27.33 -> 27.52; 26.57 -> 26.77 | Helped |
| use the wavelet filter bank instead of difference operator ∇ (ablation) | VEM-Net · Set12 | PSNR (dB): 31.75 -> 31.93; 29.59 -> 29.78; 28.29 -> 28.47; 27.31 -> 27.52; 26.55 -> 26.77 | Helped |
| train one model on Gaussian noise and apply it to Poisson-noise deblurring | VEM-Net · Set12 with Poisson noise, peaks 128–1024 | PSNR (dB): Table 5 reports Ours 25.69/26.69/27.93/29.15 and FCNN 25.09/26.09/27.27/27.95 | Helped |
Closest papers by meaning
- Variational Denoising Network: Toward Blind Noise Modeling and Removal (NeurIPS 2019 · similarity 0.90)
- Noise-Blind Image Deblurring (CVPR 2017 · similarity 0.89)
- A Variational EM Framework With Adaptive Edge Selection for Blind Motion Deblurring (CVPR 2019 · similarity 0.88)
- Deep Mean-Shift Priors for Image Restoration (NeurIPS 2017 · similarity 0.87)
- SeNM-VAE: Semi-Supervised Noise Modeling with Hierarchical Variational Autoencoder (CVPR 2024 · similarity 0.84)
- A Restoration Network as an Implicit Prior (ICLR 2024 · similarity 0.84)
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.