Research map / Image generation / Super-resolution
ZFusion: Efficient Deep Compositional Zero-shot Learning for Blind Image Super-Resolution with Generative Diffusion Prior
Alireza Esmaeilzehi, Hossein Zaredar, Yapeng Tian, Laleh Seyyed-Kalantari
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 CZSD module (degradation embeddings) (ablation) | ZFusion · RealSR ablation | CLIP-IQA: 0.6496 -> 0.6848 | Improved |
| add CZSD module (degradation embeddings) (ablation) | ZFusion · RealSR ablation | MANIQA: 0.5164 -> 0.5687 | Improved |
| add CZSD module (degradation embeddings) (ablation) | ZFusion · RealSR ablation | NIQE: 4.9553 -> 4.7152 | Improved |
| fuse synthetic and real-world degradation attributes instead of only one type (ablation) | ZFusion · RealSR ablation | MANIQA: 0.5285 -> 0.5687 | Improved |
| fuse synthetic and real-world attributes instead of only synthetic (ablation) | ZFusion · RealSR ablation | MANIQA: 0.5308 -> 0.5687 | Improved |
| add alignment loss (ablation) | ZFusion · RealSR ablation | MANIQA: 0.5258 -> 0.5687 | Improved |
| add alignment loss (ablation) | ZFusion · RealSR ablation | CLIP-IQA: 0.6609 -> 0.6848 | Improved |
| add alignment loss (ablation) | ZFusion · RealSR ablation | CNN-IQA: 0.6425 -> 0.6798 | Improved |
| use ZFusion instead of DiffBIR on synthetic degradations | ZFusion · DIV2K val synthetic degradations | MANIQA: 0.3800 -> 0.6141 | Improved |
| use ZFusion instead of DiffBIR on real-world degradations | ZFusion · RealSR real-world degradations | MANIQA: 0.3632 -> 0.5687 | Improved |
| use ZFusion instead of DiffBIR on real-world degradations | ZFusion · RealSR real-world degradations | PSNR: 22.55 -> 21.05 | Hurt |
| use ZFusion instead of DiffBIR on real-world degradations | ZFusion · RealSR real-world degradations | SSIM: 0.6240 -> 0.5788 | Hurt |
Closest papers by meaning
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator (NeurIPS 2025 · similarity 0.88)
- Navigating Beyond Dropout: An Intriguing Solution towards Generalizable Image Super Resolution (CVPR 2024 · similarity 0.88)
- One-Step Effective Diffusion Network for Real-World Image Super-Resolution (NeurIPS 2024 · similarity 0.87)
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · similarity 0.87)
- Adversarial Diffusion Compression for Real-World Image Super-Resolution (CVPR 2025 · similarity 0.87)
- CDFormer: When Degradation Prediction Embraces Diffusion Model for Blind Image Super-Resolution (CVPR 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.
