Research map / Image generation / Super-resolution
Learning a Single Convolutional Super-Resolution Network for Multiple Degradations
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 |
|---|---|---|---|
| Condition the network on blur-kernel and noise-level degradation maps | SRMD · Set5, Set14, BSD100, and Urban100 under multiple degradations | PSNR | Helped |
| Jointly handle denoising and super-resolution instead of cascading denoising before SR | SRMD · Set5 with noisy degradations | PSNR | Helped |
| Use SRMD trained for multiple degradations rather than a bicubic-specific model | SRMD · Bicubic ×4, Set5, Set14, and BSD100 | PSNR: 31.59 vs 32.05; 28.15 vs 28.49; 27.34 vs 27.58 | Mixed |
| Remove the noise-level map connection and fine-tune for noise-free degradation (ablation) | SRMD · Set5 ×2 bicubic, noise level 0 | PSNR: 37.53 -> 37.79 | Helped |
| Use degradation-matched SRMD rather than bicubic-trained VDSR | SRMD · Set5, kernel width 1.3, noise level 15, ×2 | PSNR: 24.82 vs 30.98 | Helped |
| Use non-blind degradation-conditioned SR instead of a blind model | SRMD · Complex blur degradations and real images | visual quality | Helped |
| Provide spatially varying blur-kernel and noise-level maps | SRMD · Spatially variant degradation | visual quality | Helped |
Closest papers by meaning
- Learning Controllable Degradation for Real-World Super-Resolution via Constrained Flows (ICML 2023 · similarity 0.92)
- Unified Dynamic Convolutional Network for Super-Resolution With Variational Degradations (CVPR 2020 · similarity 0.92)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.90)
- Low-Res Leads the Way: Improving Generalization for Super-Resolution by Self-Supervised Learning (CVPR 2024 · similarity 0.89)
- Unsupervised Degradation Representation Aware Transform for Real-World Blind Image Super-Resolution (AAAI 2025 · similarity 0.89)
- ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024 · 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.