Research map / Image generation / Image restoration
Neumann Network With Recursive Kernels for Single Image Defocus 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 |
|---|---|---|---|
| Use NRKNet rather than the strongest listed baseline | DPDD-trained SIDD models · DPDD test | PSNR: 25.468 -> 26.109 | Helped |
| Use NRKNet rather than the strongest listed baseline | DPDD-trained SIDD models · RealDOF | PSNR: 24.884 -> 25.148 | Helped |
| Use NRKNet rather than the strongest listed baseline | DPDD-trained SIDD models · RTF | PSNR: 24.972 -> 25.931 | Helped |
| Use NRKNet rather than the strongest listed baseline | LFDOF-trained SIDD models · LFDOF test | PSNR: 30.253 -> 30.481 | Helped |
| Use NRKNet rather than the strongest listed baseline | LFDOF-trained SIDD models · RTF | PSNR: 27.552 -> 28.047 | Helped |
| Learn atom kernels instead of using fixed Gaussian atoms (ablation) | RKR in NRKNet · DPDD | PSNR: 25.309 -> 26.109 | Helped |
| Use recursive rather than independent atoms (ablation) | RKR in NRKNet · DPDD | PSNR: 25.994 -> 26.109 | Helped |
| Use separable rather than non-separable atoms (ablation) | RKR in NRKNet · DPDD | PSNR: 26.071 -> 26.109 | Helped |
| Use cross-scale fusion rather than full Neumann expansion at every scale (ablation) | NRKNet pipeline · DPDD | PSNR: 26.077 -> 26.109 | Helped |
| Use three scales instead of two (ablation) | NRKNet pipeline · DPDD | PSNR: 25.894 -> 26.109 | Helped |
| Use three scales instead of four (ablation) | NRKNet pipeline · DPDD | PSNR: 26.020 -> 26.109 | Helped |
| Add ConvLSTM units across scales (ablation) | NRKNet coefficient prediction · DPDD | PSNR: 24.371 -> 26.109 | Helped |
Closest papers by meaning
- Prior and Prediction Inverse Kernel Transformer for Single Image Defocus Deblurring (AAAI 2024 · similarity 0.92)
- Residual Diffusion Deblurring Model for Single Image Defocus Deblurring (AAAI 2025 · similarity 0.89)
- Multi-Focus Image Fusion via Explicit Defocus Blur Modelling (AAAI 2025 · similarity 0.88)
- DeblurDiff: Real-Word Image Deblurring with Generative Diffusion Models (NeurIPS 2025 · similarity 0.87)
- Parameterized Blur Kernel Prior Learning for Local Motion Deblurring (CVPR 2025 · similarity 0.87)
- Quad-Pixel Image Defocus Deblurring: A New Benchmark and Model (CVPR 2025 · 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.