Research map / Image generation / Super-resolution
MFSR: Multi-Fractal Feature for Super-Resolution Reconstruction With Fine Details Recovery
Lianping Yang, Peng Jiao, Jinshan Pan, Hegui Zhu, Su 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 sub-denoiser (ablation) | MFSR · FFHQ 4x SR | PSNR: 26.88 -> 26.94 | Improved |
| add sub-denoiser (ablation) | MFSR · FFHQ 4x SR | FID: 70.36 -> 69.64 | Improved |
| add MFB, DEB and GPB (ablation) | MFSR · FFHQ 4x SR | PSNR: 26.79 -> 26.94 | Improved |
| add MFB, DEB and GPB (ablation) | MFSR · FFHQ 4x SR | SSIM: 0.820 -> 0.833 | Improved |
| add MFB (fractal texture prior) (ablation) | MFSR · DIV2K 4x SR | PSNR: 27.98 -> 28.16 | Improved |
| add MFB (fractal texture prior) (ablation) | MFSR · DIV2K 4x SR | FID: 108.14 -> 106.79 | Improved |
| use fractal computation instead of pure convolution texture prior (ablation) | MFSR · DIV2K 4x SR | PSNR: 28.02 -> 28.16 | Improved |
| use fractal computation instead of pure convolution texture prior (ablation) | MFSR · DIV2K 4x SR | FID: 109.70 -> 106.79 | Improved |
| increase total diffusion time steps from 50 to 1000 (ablation) | MFSR · FFHQ 4x SR | PSNR: 24.53 -> 26.94 | Improved |
| increase total diffusion time steps from 1000 to 2000 (ablation) | MFSR · FFHQ 4x SR | PSNR: 26.94 -> 27.06 | Improved |
| increase clustering kernels from 64 to 256 (ablation) | MFSR · FFHQ 4x SR | PSNR: 26.94 -> 26.76 | Hurt |
| increase MFB channels from 3 to 15 (ablation) | MFSR · FFHQ 4x SR | PSNR: 26.94 -> 26.95 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- ResDiff — ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024) · better on 12 of 12 shared results
- ResShift — ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting (NeurIPS 2023) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024 · similarity 0.86)
- Arbitrary-Scale Image Generation and Upsampling using Latent Diffusion Model and Implicit Neural Decoder (CVPR 2024 · similarity 0.86)
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · similarity 0.85)
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator (NeurIPS 2025 · similarity 0.85)
- Noise-Started One-Step Image Super-Resolution via LR-Conditioned SplitMeanFlow and GAN Refinement (NeurIPS 2026 · similarity 0.85)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.84)
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.
