Research map / Image generation / Super-resolution
Learning a Single Network for Scale-Arbitrary Super-Resolution
Longguang Wang, Yingqian Wang, Zaiping Lin, Jungang Yang, Wei An, Yulan Guo
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 scale-aware upsampling instead of bicubic upsampling (ablation) | EDSR · Set5 x1.7/x2 | PSNR: 39.56/37.86 -> 39.76/38.13 | Improved |
| add scale-aware convolution to feature adaption (ablation) | variant 2 · Set5 x1.3/x3.9 | PSNR: 34.32 -> 34.38 | Improved |
| add guidance map generation (ablation) | variant 3 · Set5 x1.3/x3.9 | PSNR: 34.38 -> 34.44 | Improved |
| add scale-aware feature adaption (conv + guidance map) (ablation) | variant 2 · Set5 x1.7/x2/x1.9/x3.5 | PSNR: 39.76/38.13/34.91 -> 39.87/38.19/35.03 | Improved |
| use 4 experts instead of 1 in scale-aware convolution (ablation) | ArbEDSR · Set5 x1.7/x2/x2.55 | PSNR: 39.37/37.87/35.82 -> 39.88/38.22/36.03 | Improved |
| use 4 experts instead of 1 on highly asymmetric scale (ablation) | ArbEDSR · Set5 x1.3/x3.9 | PSNR: 33.34 -> 34.44 | Improved |
| increase experts from 4 to 8 (ablation) | ArbEDSR · Set5 x1.3/x3.9 | PSNR: 34.44 -> 34.46 | No clear effect |
| plug scale-aware module into EDSR vs bicubic-resized EDSR | ArbEDSR · Set5 x1.6/x1.55 | PSNR: 40.39/40.71 -> 40.64/40.94 | Improved |
| plug scale-aware module into EDSR at integer scale x2 | ArbEDSR · Set5 x2 | PSNR: 38.19 -> 38.19 | No clear effect |
| replace Meta-RCAN with ArbRCAN (scale-aware conv) on asymmetric SR | Meta-RCAN+ft · Manga109 asymmetric | PSNR: 37.80/33.67/34.28 -> 37.93/33.81/34.41 | Improved |
| fine-tune Meta-RCAN on the authors' training set | Meta-RCAN · Manga109 asymmetric | PSNR: 37.74/33.61/34.23 -> 37.80/33.67/34.28 | No clear effect |
| use ArbRCAN instead of Meta-RCAN for memory and time | Meta-RCAN · B100/LR 100x100 x4 | memory, time: 3.1G -> 1.1G; 0.39s -> 0.29s | Improved |
Closest papers by meaning
- SAVSR: Arbitrary-Scale Video Super-Resolution via a Learned Scale-Adaptive Network (AAAI 2024 · similarity 0.91)
- Meta-SR: A Magnification-Arbitrary Network for Super-Resolution (CVPR 2019 · similarity 0.90)
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks (CVPR 2016 · similarity 0.90)
- Deep Arbitrary-Scale Image Super-Resolution via Scale-Equivariance Pursuit (CVPR 2023 · similarity 0.90)
- Adaptive Importance Learning for Improving Lightweight Image Super-Resolution Network (IJCV 2019 · similarity 0.89)
- FSR: A General Frequency-Oriented Framework to Accelerate Image Super-resolution Networks (AAAI 2023 · 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.
