Research map / Image generation / Super-resolution
Learning Correction Filter via Degradation-Adaptive Regression for Blind Single Image Super-Resolution
Hongyang Zhou, Xiaobin Zhu, Jianqing Zhu, Zheng Han, Shi-Xue Zhang, Jingyan Qin, Xu-Cheng Yin
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 metric loss (DM loss) (ablation) | Ours + SRFBN-s · NTIRE2017 Track2 x4 | PSNR: +1.29 dB PSNR, +0.1367 LPIPS gain as stated | Improved |
| use Gaussian + DoG dictionary instead of RAISR filters (ablation) | DARM · NTIRE2017 Track2 x4 | PSNR: 22.12 -> 23.35 | Improved |
| use Gaussian + DoG dictionary instead of directly learned filters (ablation) | DARM · NTIRE2017 Track2 x4 | PSNR: 22.44 -> 23.35 | Improved |
| increase dictionary size from 24 to 72 Gaussian + DoG filters (ablation) | DARM · NTIRE2017 Track2 x4 | PSNR: 23.17 -> 23.35 | Improved |
| train the base SR model with Gaussian or Lanczos degradation instead of bicubic (ablation) | Ours + SRFBN-s · NTIRE2018 Track2 x4 | PSNR: 20.32 -> 20.31 (Gaussian); 20.32 (Lanczos) | No clear effect |
| apply the learned correction filter in front of SRFBN-s | SRFBN-s · NTIRE2017 Track2 x4 | PSNR: 23.07 -> 23.35 | Improved |
| apply the learned correction filter in front of SRFBN-s | SRFBN-s · NTIRE2017 Track2 x4 | LPIPS: 0.4845 -> 0.5063 | Hurt |
| use the prior Correction Filter method instead of DARM in front of SR models | SRFBN-s · NTIRE2017 Track2 x4 | PSNR: 23.07 -> 20.98 | Hurt |
Closest papers by meaning
- Unsupervised Degradation Representation Learning for Blind Super-Resolution (CVPR 2021 · similarity 0.93)
- AND: Adversarial Neural Degradation for Learning Blind Image Super-Resolution (NeurIPS 2023 · similarity 0.92)
- End-to-End Alternating Optimization for Real-World Blind Super Resolution (IJCV 2023 · similarity 0.92)
- Unsupervised Degradation Representation Aware Transform for Real-World Blind Image Super-Resolution (AAAI 2025 · similarity 0.90)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.89)
- Navigating Beyond Dropout: An Intriguing Solution towards Generalizable Image Super Resolution (CVPR 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.
