Reproduce “Diffusion Models for Image Restoration and Enhancement: A Comprehensive Survey” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Xin Li, Yulin Ren, Xin Jin, Cuiling Lan, Xingrui Wang, Wenjun Zeng, Xinchao Wang, Zhibo Chen

The authors survey diffusion models for image restoration, organizing methods as supervised or zero-shot and grouping blind-restoration strategies. Their comparison shows a metric trade-off on 4× DIV2K super-resolution: diffusion method ResShift has higher PSNR than GAN-based SeD (29.38 versus 29.27 dB), while SeD has lower LPIPS and FID. This compares selected methods, not general superiority.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.