LYON19- Lymphocyte Detection Test Set

<strong>LYON19</strong> The provided test set includes 441 ROIs saved in the .<em>png</em> files, and it is a test set of LYON grand challenge: https://lyon19.grand-challenge.org <strong>Data Description</strong> The test set contains Region of Interests (ROIs) selected from whole-slide images (WSI) of immunohistochemistry (IHC) stained specimens of breast, colon and prostate. Data came from eight different medical centers in the Netherlands. All slides were stained with an antibody against CD3 or CD8. Slides were subsequently digitized with a Pannoramic 250Flash II scanner (3DHistech, Hungary), resulting in WSIs with a spatial resolution of 0.24μm/px. Selected ROIs were saved with full resolution in the .<em>png</em> files. Selected ROIs were representative for most different types of lymphocyte distributions that occur in slides, namely (1) area with regular lymphocyte distribution, (2) clustered cells, and (3) staining or tissue artifacts. <strong>Citation:</strong> Please reference the following paper if you use LYON19 data for a scientific publication: Swiderska-Chadaj, Zaneta, et al. "<em><strong>Learning to detect lymphocytes in immunohistochemistry with deep learning</strong></em>." Medical Image Analysis (2019): 101547. Link to the paper: https://www.sciencedirect.com/science/article/pii/S1361841519300829<strong>LYON19</strong> 本测试集共包含441个感兴趣区域（Region of Interests，ROIs），均存储于<em>png</em>格式文件中，该数据集为LYON19 grand challenge的官方测试集，官网链接：https://lyon19.grand-challenge.org <strong>数据说明</strong> 本测试集的感兴趣区域（ROIs）提取自乳腺、结肠与前列腺组织的免疫组织化学（Immunohistochemistry，IHC）染色标本的全切片图像（Whole-slide Images，WSIs）。该数据集采集自荷兰境内8家不同的医疗中心。所有切片均使用针对CD3或CD8的抗体完成染色。随后，使用匈牙利3DHistech公司生产的Pannoramic 250Flash II扫描仪对切片进行数字化处理，得到空间分辨率为0.24μm/像素的全切片图像。所选取的ROIs以全分辨率存储于<em>png</em>格式文件中，且能够代表切片中出现的绝大多数淋巴细胞分布类型，具体包括：(1) 淋巴细胞均匀分布区域，(2) 细胞聚集体区域，以及(3) 染色伪影或组织伪影区域。 <strong>引用说明</strong> 若您在学术出版物中使用LYON19数据集，请引用以下文献：Swiderska-Chadaj, Zaneta 等人发表于《Medical Image Analysis》2019年的论文<em><strong>基于深度学习检测免疫组织化学图像中的淋巴细胞</strong></em>，文献编号：101547。论文链接：https://www.sciencedirect.com/science/article/pii/S1361841519300829