MLAGNet: Multi-task Medical Image Datasets (Super-Resolution, Segmentation, Classification)

Datasets for the MLAGNet code repository (https://github.com/JGuan-lab/MLAGNet), a unified feature-extraction framework for medical imaging covering super-resolution, segmentation and classification.Files• superresolution.tar — lung CT (lungtotal=HR ground truth, lungdegrade=degraded), cropped CT (qingtcroplow/qingtcroptrue) and Kaggle COVID CT at multiple scales (kagglecovid).• segmentation.tar — angiography, crack and retinal-vessel (retinaplus) datasets, each with images + masks (labelcol).• classification.tar.part00 ... part08 — the classification archive (MHIST, Chest X-Ray, ISIC) split into 2 GB parts because Zenodo rejects very large single files. To reassemble: cat classification.tar.part* > classification.tar && tar -xf classification.tarThe breast histopathology classification dataset is not included here due to size; it is available at https://huggingface.co/datasets/EulerianKnight/breast-histopathology-images-train-test-valid-split .