Research map / Image generation / Super-resolution
Efficient and Degradation-Adaptive Network for Real-World Image Super-Resolution
Jie Liang, Hui Zeng, Lei Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add degradation-adaptive mixture of experts to SRResNet | SRResNet · DIV2K val Bicubic | PSNR (dB, Y): 28.05 -> 28.55 | Improved |
| add degradation-adaptive mixture of experts to SRResNet | SRResNet · DIV2K val Level-II | LPIPS: 0.2228 -> 0.2126 | Improved |
| add degradation-adaptive mixture of experts to SRResNet | SRResNet · DIV2K val Level-III | PSNR (dB, Y): 23.71 -> 23.93 | Improved |
| DASR versus Real-ESRGAN on Level-I | DASR · DIV2K val Level-I | PSNR (dB, Y): 26.17 -> 27.84 | Improved |
| DASR versus BSRGAN on Level-III | DASR · DIV2K val Level-III | LPIPS: 0.3995 -> 0.4144 | Hurt |
| use DASR backbone instead of heavy RRDB | DASR · DIV2K val Bicubic | PSNR (dB, Y): 27.92 -> 28.55 | Improved |
| use DASR instead of RRDB on perceptual metric | DASR · DIV2K val Bicubic | LPIPS: 0.1473 -> 0.1696 | Hurt |
| use 9 experts instead of 5 (ablation) | DASR · real-world images (visual) | visual quality | No clear effect |
| use 3 experts instead of 5 (ablation) | DASR · real-world images (visual) | visual quality | Hurt |
| add sigmoid layer to the weighting module (ablation) | DASR · real-world images (visual) | visual quality | No clear effect |
| mix expert outputs as in classical MoE instead of mixing parameters (ablation) | DASR · real-world images (visual) | visual quality | Hurt |
| replace joint multi-expert mixture with other dynamic convolutions (ablation) | DASR · real-world images (visual) | visual quality | Hurt |
Closest papers by meaning
- Learning a Single Convolutional Super-Resolution Network for Multiple Degradations (CVPR 2018 · similarity 0.92)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.91)
- Unsupervised Degradation Representation Learning for Blind Super-Resolution (CVPR 2021 · similarity 0.91)
- Unsupervised Degradation Representation Aware Transform for Real-World Blind Image Super-Resolution (AAAI 2025 · similarity 0.89)
- Low-Res Leads the Way: Improving Generalization for Super-Resolution by Self-Supervised Learning (CVPR 2024 · similarity 0.89)
- One-Step Effective Diffusion Network for Real-World Image Super-Resolution (NeurIPS 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.
