PanNuke
收藏资源简介:
PanNuke是一个半自动生成的细胞核实例分割和分类数据集,包含19种不同组织类型的详尽细胞核标签。数据集总共有189,744个标记的细胞核,每个细胞核都有一个实例分割掩码。数据集分为三个折叠(fold1, fold2, fold3),每个折叠包含图像、实例、类别和组织类型四个特征。图像为RGB格式,实例为二进制格式的细胞核实例,类别为细胞核的类别标签,组织类型为样本来源的组织类型标签。
PanNuke is a semi-automatically generated dataset for nuclear instance segmentation and classification, featuring comprehensive nuclear annotations across 19 distinct tissue types. The dataset contains a total of 189,744 labeled nuclei, each paired with an instance segmentation mask. It is split into three folds (fold1, fold2, fold3), where each fold includes four feature categories: images, instances, category labels, and tissue type labels. The images are stored in RGB format, the instances are binary-format nuclear instance masks, the category labels denote the class of the corresponding nucleus, and the tissue type labels indicate the tissue origin of the source sample.




