Reproduce “PixelSR: Pixel-Space Orthogonal Flow Decoupling for Real-World Image Super-Resolution” 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
Xiang Lv, Kai Zhao, Qiang Wang, Qiao Mo, Chao Zhou, Mingwen Shao
PixelSR restores real-world low-resolution images directly in RGB, separating structure and texture prediction within flow matching. Its orthogonal flow decomposition assigns projected and complementary targets to different branches. On LSDIR, against monolithic prediction with the same network and parameter count, it reports PSNR 20.6907 and MANIQA 0.5661. It ranks first by average rank on four datasets, but not every metric.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.