Research map / Image generation / Super-resolution
Learning the Non-Differentiable Optimization for Blind Super-Resolution
Check 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 AMNet L-RL for blind SR instead of DAN | blind SR · Set5, Gaussian8, ×4 | PSNR: 31.89 -> 32.34 | Helped |
| use AMNet L-RL for blind SR instead of DAN | blind SR · Set14, Gaussian8, ×4 | PSNR: 28.43 -> 28.57 | Helped |
| use AMNet L-RL for blind SR instead of DAN | blind SR · BSD100, Gaussian8, ×4 | PSNR: 27.51 -> 27.66 | Helped |
| use AMNet L-RL for blind SR instead of DAN | blind SR · Urban100, Gaussian8, ×4 | PSNR: 25.86 -> 25.86 | No clear effect |
| use AMNet L-RL for blind SR instead of DAN | blind SR · Manga109, Gaussian8, ×4 | PSNR: 30.50 -> 30.39 | Hurt |
| use RL to select the blur-kernel code for AMGAN | AMGAN blind SR · Set5, kernel widths 0.2, 1.3 and 2.6 | NIQE: 5.047 -> 4.8206 | Helped |
| use RL to select the blur-kernel code for AMGAN | AMGAN blind SR · Set14, kernel widths 0.2, 1.3 and 2.6 | NIQE: 4.2598 -> 3.8897 | Helped |
| use RL to select the blur-kernel code for AMGAN | AMGAN blind SR · BSD100, kernel widths 0.2, 1.3 and 2.6 | NIQE: 4.0750 -> 3.4968 | Helped |
| use RL to select the blur-kernel code for AMGAN | AMGAN blind SR · PIRM Val, kernel widths 0.2, 1.3 and 2.6 | NIQE: 3.3672 -> 3.0021 | Helped |
| replace AMLayer with AdaIN (ablation) | AMNet · Set5, Set14, BSD100 and Urban100, blur kernel width 1.8 | PSNR: 32.12 -> 32.43; 28.53 -> 28.67; 27.56 -> 27.68; 25.66 -> 25.81 | Hurt |
| remove the different-sign penalty from the critic loss (ablation) | AMGAN-RL · Set5, kernel widths 0.2, 1.3 and 2.6 | NIQE: 4.88 / 4.77 / 4.80 -> 4.98 / 4.83 / 4.87 | Hurt |
Closest papers by meaning
- Blind Super-Resolution With Iterative Kernel Correction (CVPR 2019 · similarity 0.90)
- Unfolding the Alternating Optimization for Blind Super Resolution (NeurIPS 2020 · similarity 0.90)
- Navigating Beyond Dropout: An Intriguing Solution towards Generalizable Image Super Resolution (CVPR 2024 · similarity 0.88)
- Augmenting Perceptual Super-Resolution via Image Quality Predictors (CVPR 2025 · similarity 0.86)
- A Dynamic Kernel Prior Model for Unsupervised Blind Image Super-Resolution (CVPR 2024 · similarity 0.86)
- SAVSR: Arbitrary-Scale Video Super-Resolution via a Learned Scale-Adaptive Network (AAAI 2024 · 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.