reproducible-image-denoising-state-of-the-art

Why we included this project

Anyone comparing single-image denoising methods will find this collection saves the usual digging through papers. It groups the well-known algorithms by family, so filtering approaches like BM3D and NLM sit alongside sparse coding methods such as KSVD and LSSC, classical priors like EPLL, low-rank techniques such as WNNM, and deep learning entries including DnCNN and RED. That layout makes it easy to see how the field moved from handcrafted priors to learned networks. The real selling point is the rule that every listed work must ship working code and reproducible results, which means the entries point straight to papers, project pages, and implementations you can actually run. It is a curated bibliography rather than ready-to-use software, and for someone assembling a denoising benchmark suite, that is exactly the useful part.