AMOS

该数据集是一个大型的临床数据集，包含500个磁共振成像（MRI）和100个计算机断层扫描（CT）的资料，其中对15个腹部器官进行了标注。为了获得统一的切片尺寸，每个切片都被填充至512×512的最终尺寸。在少样本实验中，我们分别使用了500、1000、2000、5000和7000个切片。具体来说，这个数据集用于训练的切片总数为10,300张，其任务是进行医学图像分割。

This dataset is a large-scale clinical dataset containing 500 magnetic resonance imaging (MRI) scans and 100 computed tomography (CT) scans, with annotations for 15 abdominal organs. To achieve uniform slice dimensions, each slice was padded to a final size of 512×512. For few-shot experiments, we utilized 500, 1000, 2000, 5000 and 7000 slices respectively. Specifically, the total number of training slices for this dataset is 10,300, and the task is medical image segmentation.