CC-19

目前，我们介绍了一个与最新的冠状病毒家族相关的小型新数据集，即 COVID-19。这些数据集在临床医学相关应用的人工智能领域发挥着重要作用。该数据集包含 89 名受试者的计算机断层扫描 (CT) 切片。在这 89 名受试者中，68 名确诊为 COVID-19 病毒患者（阳性病例），其余 21 名被发现为阴性病例。提议的数据集“CC-19”包含属于 98 名受试者的 34,006 个 CT 扫描切片（图像），其中 28,395 个 CT 扫描切片属于 COVID 阳性患者。该数据集是公开的。第一个图显示了从 CC-19 数据集的 CT 扫描中获取的一些 2D 切片。此外，从数据集中选择的一些 3D 样本如图所示。亨斯菲尔德单位 (HU) 是 CT 扫描辐射密度的测量值，如表所示。通常，CT 扫描设备经过仔细校准以测量 HU 单位。该单元可用于提取CT扫描切片中的相关信息。 CT 扫描切片具有圆柱形扫描边界。由于未知原因，位于此圆柱边界之外的像素信息被 CT 扫描仪系统自动丢弃。但幸运的是，这种外部像素的丢弃消除了一些预处理步骤。

Currently, we introduce a novel small-scale dataset associated with the latest coronavirus family, specifically for COVID-19. Such datasets play a critical role in the field of artificial intelligence for clinical medicine-related applications. This dataset contains computed tomography (CT) slices from 89 subjects. Among these 89 subjects, 68 were confirmed as COVID-19 positive patients, while the remaining 21 were identified as negative cases. The proposed dataset "CC-19" includes 34,006 CT scan slices (images) belonging to 98 subjects, of which 28,395 CT scan slices belong to COVID-positive patients. This dataset is publicly available. The first figure displays some 2D slices extracted from the CT scans of the "CC-19" dataset. Additionally, several selected 3D samples from the dataset are illustrated in the figures. The Hounsfield Unit (HU) is a measure of the radiodensity of CT scans, as shown in the table. Typically, CT scanning equipment is meticulously calibrated to measure HU units. This unit can be utilized to extract relevant information from CT scan slices. CT scan slices feature a cylindrical scanning boundary. For unknown reasons, pixel information located outside this cylindrical boundary is automatically discarded by the CT scanner system. Fortunately, the discarding of these external pixels eliminates some preprocessing steps.

搜集汇总

背景与挑战

背景概述

CC-19是一个包含98名受试者34,006个CT扫描切片的数据集，主要用于COVID-19检测研究，其中28,395个切片来自阳性患者。该数据集由电子科技大学于2021年发布，支持临床医学AI应用。