Federated Tumor Segmentation (FeTS) 2022 Challenge Dataset

该数据集包含了来自1251名多机构患者的训练数据和219例外部验证案例，用于评估肿瘤分割的性能。在联邦学习的背景下，该数据集包含了如二进制DICE相似性和豪斯多夫距离等性能评价指标。规模上，数据集涵盖了1251名训练患者和219个验证案例。任务方面，数据集旨在进行总肿瘤、肿瘤核心和增强肿瘤的语义分割。

This dataset contains training data from 1251 multi-institutional patients and 219 external validation cases, which is utilized for assessing the performance of tumor segmentation. In the context of federated learning, this dataset provides performance evaluation metrics including binary DICE similarity and Hausdorff distance. Regarding its scale, the dataset covers 1251 training patients and 219 validation cases. The primary task of this dataset is semantic segmentation of whole tumor, tumor core, and enhancing tumor.