FHIST

FHIST数据集是一个专为少样本组织学图像分类设计的公开基准。由蒙特利尔高等技术学院的研究团队创建，该数据集汇集了来自多个公共数据集的多样化样本，旨在解决医学成像中标记数据稀缺的问题。数据集包含280,000个样本，覆盖了多种组织类型和不同级别的域转移，反映了现实场景中的组织分类任务。FHIST数据集的应用领域包括癌症相关组织分类，旨在通过少量的标记样本来训练模型，以适应新的分类任务。

The FHIST dataset is a publicly available benchmark tailored for few-shot histological image classification. Developed by a research team at the École de technologie supérieure (ÉTS) of Montreal, this dataset aggregates diverse samples from multiple public datasets to address the critical issue of scarce labeled data in medical imaging. Comprising 280,000 samples, it covers a wide spectrum of tissue types and domain shifts across varying levels, which accurately reflects real-world tissue classification scenarios. The FHIST dataset finds applications in cancer-related tissue classification, enabling model training with only a small number of labeled samples to adapt to novel classification tasks.