Research map / Image generation / Super-resolution
Deep Plug-And-Play Super-Resolution for Arbitrary Blur Kernels
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 DPSR instead of IRCNN+RCAN | DPSR · BSD68, ×2, Gaussian, motion, and disk blur; ×3 with noise; ×4 blur settings | PSNR: 27.28 vs 24.79; 30.05 vs 28.54; 28.61 vs 25.48; 25.22 vs 23.52; 27.22 vs 25.88; 26.19 vs 23.82; 23.92 vs 22.95; 24.94 vs | Helped |
| Add explicit blur handling through the FFT data-fidelity update and iterate with a super-resolver prior | DPSR · BSD68 synthetic arbitrary blur kernels | PSNR: 27.28 vs 24.79 | Helped |
| Use bicubic-trained SRResNet+ as the prior within DPSR rather than SRResNet+ alone | DPSR · BSD68, ×2, Gaussian blur, noise 0 | PSNR: 27.28 | Helped |
| Use SRResNet+ rather than SRResNet | SRResNet+ · Set5, bicubic degradation | PSNR: +0.15dB | Helped |
| Apply a bicubic-trained model directly to blurry LR images without explicit blur handling | VDSR and RCAN · Synthetic arbitrary blur kernels | PSNR: 23.47 vs 27.28 | Hurt |
| Use a deblurring-then-SISR cascade instead of joint iterative DPSR | IRCNN+RCAN · Synthetic arbitrary blur kernels | PSNR: 24.79 vs 27.28 | Hurt |
| Use ZSSR for large complex blur kernels | ZSSR · Synthetic arbitrary blur kernels | PSNR | Hurt |
| Use DeblurGAN+RCAN or GFN for large complex blur rather than DPSR | DeblurGAN+RCAN and GFN · Synthetic arbitrary blur kernels | PSNR | Hurt |
| Use SRGAN+ as the plug-in super-resolver prior | DPSRGAN · Scale factor 4 visual comparison | visual quality | Helped |
Closest papers by meaning
- Learning a Single Convolutional Super-Resolution Network for Multiple Degradations (CVPR 2018 · similarity 0.91)
- Spectrum-to-Kernel Translation for Accurate Blind Image Super-Resolution (NeurIPS 2021 · similarity 0.90)
- Unsupervised Degradation Representation Aware Transform for Real-World Blind Image Super-Resolution (AAAI 2025 · similarity 0.87)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.87)
- A Dynamic Kernel Prior Model for Unsupervised Blind Image Super-Resolution (CVPR 2024 · similarity 0.87)
- ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024 · similarity 0.87)
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.