DeepDRiD

该数据集是一个公开可用的眼底成像数据集，包含了五个等级的分类，遵循MedMnist的划分方式，其中训练集有1,080个样本，验证集有120个样本，测试集有400个样本。此外，该数据集在训练、验证和测试方面遵循MedMnist的划分方法。整个数据集规模为1,600个样本，任务是对眼底图像进行分类。

This is a publicly available fundus imaging dataset that supports five-class classification, following the data partitioning scheme of MedMNIST. The dataset splits into 1,080 training samples, 120 validation samples, and 400 test samples. The total number of samples across the entire dataset is 1,600, and the core task of this dataset is fundus image classification. Additionally, the training, validation, and test splits all adhere to the MedMNIST partitioning methodology.