Synthetic and Real Rim Lesions

该数据集包含了260个真实的边缘病变样本和7720个非边缘病变样本，同时为了提升分类器的性能，还加入了人工生成的边缘病变样本。此外，数据集还包括了合成的定量磁化率映射和相关概率分割映射，旨在提高对罕见顺磁性边缘病变的检测和分类能力。总体而言，该数据集共有7980个样本，涵盖了合成和真实样本，任务是对磁共振成像中的边缘病变进行分类和分割。

This dataset contains 260 real borderline lesion samples and 7720 non-borderline lesion samples. Additionally, artificially generated borderline lesion samples were added to improve the performance of the classifier. Furthermore, the dataset also includes synthetic quantitative susceptibility maps and corresponding probabilistic segmentation maps, which aim to enhance the detection and classification capabilities for rare paramagnetic borderline lesions. Overall, this dataset has a total of 7980 samples covering both synthetic and real specimens, with the task of classifying and segmenting borderline lesions in magnetic resonance imaging (MRI).