Reproduce “Toward Convolutional Blind Denoising of Real Photographs” 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
Shi Guo, Zifei Yan, Kai Zhang, Wangmeng Zuo, Lei Zhang
CBDNet addresses real-camera photo denoising by modeling signal-dependent noise and camera processing, then estimating a spatial noise map to guide cleanup. It trains on mixed examples, and the map can be adjusted to balance noise removal against detail. On DND’s 50-image sRGB benchmark, CBDNet(All) reports 38.06 dB PSNR and 0.9421 SSIM, narrowly above blind denoiser TWSC; this is a small comparison-specific lead.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.