Research map / Image generation / Super-resolution
Spectrum-to-Kernel Translation for Accurate Blind Image Super-Resolution
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 |
|---|---|---|---|
| Use S2K kernel estimation with RCAN rather than IKC | Blind SR · DIV2K, Gaussian, 2× | PSNR: 28.45 -> 29.77 | Helped |
| Use S2K kernel estimation with RCAN rather than IKC | Blind SR · Flicker2K, Gaussian, 4× | PSNR: 25.64 -> 26.10 | Helped |
| Use S2K kernel estimation with SFTMD rather than IKC | Blind SR · DIV2K, Motion, 2× | PSNR: 24.96 -> 30.95 | Helped |
| Add the relative shape loss L_s (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | PSNR: 19.46 -> 26.68 | Helped |
| Add the relative shape loss L_s (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | SSIM: 0.4952 -> 0.7562 | Helped |
| Add the smoothing loss L_reg (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | PSNR: 26.68 -> 26.94 | Helped |
| Add the smoothing loss L_reg (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | SSIM: 0.7562 -> 0.7563 | Helped |
| Add the pixel loss L_1 (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | PSNR: 8.462 -> 26.68 | Helped |
| Use DFT spectrum input instead of spatial input, with kernel-matrix output (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | PSNR: 22.50 -> 26.96 | Helped |
| Predict an up-sampled kernel matrix instead of kernel parameters (ablation) | S2K with DFT input · DIV2K validation, random Gaussian kernels, 3× | PSNR: 20.53 -> 26.96 | Helped |
| Use DFT spectrum input and up-sampled kernel-matrix output (ablation) | S2K · DIV2K validation, random Gaussian kernels, 3× | PSNR: 20.01 -> 26.96 | Helped |
Closest papers by meaning
- Blind Super-Resolution With Iterative Kernel Correction (CVPR 2019 · similarity 0.93)
- Deep Constrained Least Squares for Blind Image Super-Resolution (CVPR 2022 · similarity 0.91)
- Unfolding the Alternating Optimization for Blind Super Resolution (NeurIPS 2020 · similarity 0.91)
- A Dynamic Kernel Prior Model for Unsupervised Blind Image Super-Resolution (CVPR 2024 · similarity 0.89)
- Unsupervised Degradation Representation Aware Transform for Real-World Blind Image Super-Resolution (AAAI 2025 · similarity 0.87)
- Navigating Beyond Dropout: An Intriguing Solution towards Generalizable Image Super Resolution (CVPR 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.