MedOtter/PanNuke

PanNuke Pan-Cancer H&E Nuclei Instance Segmentation and Classification dataset (Gamper et al., ECDP 2019; arXiv:2003.10778). Mirrored from the Warwick TIA Centre release. Composition 7,901 RGB patches of 256x256 at 40x magnification (~0.25 um/pixel) 19 tissue types pooled from TCGA / GTEx (Breast, Colon, Lung, Kidney, Prostate, Stomach, Ovarian, Bladder, Esophagus, Pancreatic, Thyroid, Skin, Cervix, Adrenal_gland, Bile-duct, Liver, HeadNeck, Testis, Uterus) ~189… See the full description on the dataset page: https://huggingface.co/datasets/MedOtter/PanNuke.