US30K

该数据集是一个全面性的超声图像集合，包含了约30,000张图片和69,000个掩膜，覆盖了六个不同的目标类别。该数据集专门构建用于评估SAMUS模型的有效性。此外，该数据集是从七个公开可用的数据集中整合而成的，包含了用于训练、验证和测试的数据分区，确保了泛化评估的准确性。规模上，该数据集大约有30,000张图片和69,000个掩膜，其任务是进行超声图像分割。

This dataset is a comprehensive collection of ultrasound images, containing approximately 30,000 images and 69,000 masks covering six distinct target categories. It was specifically constructed to evaluate the effectiveness of the SAMUS model. Furthermore, this dataset is integrated from seven publicly available datasets, and includes data splits for training, validation, and testing to ensure the accuracy of generalization evaluation. Boasting around 30,000 images and 69,000 masks, its core task is ultrasound image segmentation.