Research map / Image generation / Super-resolution
Pixel-Aware Stable Diffusion for Realistic Image Super-Resolution and Personalized Stylization
Tao Yang, Rongyuan Wu, Peiran Ren, Xuansong Xie, Lei Zhang
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 degradation removal module (ablation) | PASD · RealSR | FID: 56.7890 -> 47.3440 FID; 0.3822 -> 0.2806 LPIPS | Improved |
| add high-level information (classification, detection, caption) (ablation) | PASD · RealSR | FID: 50.2502 -> 47.3440 | Improved |
| add negative prompt with classifier-free guidance (ablation) | PASD · RealSR | FID: 52.3380 -> 47.3440 FID; runtime 8.7366 -> 14.5889 s; PSNR 27.0901 -> 25.9301 | Mixed |
| remove pixel-aware cross attention (add ControlNet features by plain addition) | PASD · RealSR | visual structure consistency | Hurt |
| increase adjustable-noise level alpha_a (ablation) | PASD · RealSR | PSNR vs QAlign: PSNR increases, QAlign decreases as alpha_a grows | Mixed |
| use diffusion-based Real-ISR rather than GAN-based | Real-ISR · RealSR / DIV2K / DRealSR | PSNR/SSIM | Hurt |
| use PASD with swapped personalized base model for cartoonization (no extra training) | FFHQ / Flicker2K · cartoonization | FID: 37.9589 -> 37.6698 (FFHQ vs ControlNet) | Improved |
| swap base model to an aesthetic one for old photo restoration | PASD · Old photo restoration | FID: 262.1837 -> 240.2576 vs DiffBIR | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 3 on most shared results.
- SeeSR — SeeSR: Towards Semantics-Aware Real-World Image Super-Resolution (CVPR 2024) · better on 2 of 4 shared results
- ControlNet — ControlNet++: Improving Conditional Controls with Efficient Consistency Feedback (ECCV 2024) · better on 1 of 1 shared result
- InstructPix2Pix — InstructPix2Pix: Learning To Follow Image Editing Instructions (CVPR 2023) · better on 1 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
- One-Step Effective Diffusion Network for Real-World Image Super-Resolution (NeurIPS 2024 · similarity 0.88)
- Simpler Diffusion: 1.5 FID on ImageNet512 with Pixel-space Diffusion (CVPR 2025 · similarity 0.87)
- When Latents Forget Pixels: Restoring Fidelity in Diffusion Transformer Super-Resolution (NeurIPS 2026 · similarity 0.87)
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator (NeurIPS 2025 · similarity 0.87)
- Taming Diffusion Prior for Image Super-Resolution with Domain Shift SDEs (NeurIPS 2024 · similarity 0.87)
- RAP-SR: RestorAtion Prior Enhancement in Diffusion Models for Realistic Image Super-Resolution (AAAI 2025 · 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.
