Strong histopathological relevance patches for MSI vs. MSS classification

This is a subset of dataset "Histological images for MSI vs. MSS classification in gastrointestinal cancer, FFPE samples" CRC_DX[1]. For more details, please visit [1]. After a fused distillation, a neural network based framework will vote for the candidates for the histopathological strong relevance patches. We incorporate the morphology knowledge from pathologists to sort out the most representative samples and tag them to this subdata. The dataset has two labels i.e., MSIMUT_strong and MSS_strong. MSIMUT_strong : 28 307 patches. MSS_strong : 27 146 patches. [1] Histological images for MSI vs. MSS classification in gastrointestinal cancer, FFPE samples. https://doi.org/10.5281/zenodo.2530835本数据集为“胃肠道癌福尔马林固定石蜡包埋（Formalin-Fixed Paraffin-Embedded，FFPE）样本中用于微卫星不稳定（Microsatellite Instability，MSI）与微卫星稳定（Microsatellite Stable，MSS）分类的组织学图像”数据集CRC_DX的子集[1]。如需获取更多细节，请参阅参考文献[1]。经融合蒸馏流程后，基于神经网络的框架将对组织病理学强相关候选图像块进行投票遴选。我们引入病理学家的形态学知识，筛选出最具代表性的样本并为该子数据集完成标注。该子数据集包含两类标签：强MSI突变（MSIMUT_strong）与强微卫星稳定（MSS_strong）。其中强MSI突变样本共28307个图像块，强微卫星稳定样本共27146个图像块。[1] Histological images for MSI vs. MSS classification in gastrointestinal cancer, FFPE samples. https://doi.org/10.5281/zenodo.2530835