adlito/covid19-cxr-bm3d-denoising

该数据集包含用于监督图像去噪的配对胸部X射线（CXR）图像。它基于COVID-19放射学数据库构建，并为每个原始CXR提供：原始（干净）图像、合成噪声版本和BM3D去噪版本（伪地面真值）。目标是训练和评估能够从噪声CXR中学习重现BM3D的模型。重要提示：这是一个衍生数据集，原始图像及其版权属于COVID-19放射学数据库的作者，使用仅限于学术/非商业目的（参见许可）。This dataset contains paired chest X-ray (CXR) images for supervised image denoising. It is built from the COVID-19 Radiography Database and, for each original CXR, provides: the original (clean) image, a synthetically noisy version, a BM3D-denoised version (pseudo–ground truth). The goal is to train and evaluate models that learn to reproduce BM3D from noisy CXRs. Important: This is a derived dataset. The original images and their copyrights belong to the authors of the COVID-19 Radiography Database. Usage is restricted to academic / non-commercial purposes (see Licensing).