胸部CT器官标注数据

通过对DICOM格式的原始胸部CT影像数据进行标注，标注出影像中的以下器官组织：气管、结节、右动脉、右静脉、左动脉、左静脉。以及18个肺段：右肺尖段、右肺后段、右肺前段、右肺外段、右肺内段、右肺背段、右肺内基底段、右肺前基底段、右肺外基底段、右肺后基底段、左肺尖后段、左肺前段、左肺上舌段、左肺下舌段、左肺背段、左肺内前基底段、左肺外基底段、左肺后基底段。提供给影像自动分割人工智能模型进行训练，帮助人工智能模型更好地识别胸部CT影像中的各器官组织和病灶，提取特征，发现规律，最终提高诊断人工智能模型的准确性、鲁棒性和泛化能力。1.数据采集：通过正式合作协议，从医疗机构取得匿名化的胸部CT数据。2.数据处理：依据人体解剖学、医学影像诊断学相关知识，结合CT影像学特征，按照1人标注、2人审查、1位临床专家确认的流程，对数据进行标注，标注出以下所有部位：气管、结节、右动脉、右静脉、左动脉、左静脉、右肺尖段、右肺后段、右肺前段、右肺外段、右肺内段、右肺背段、右肺内基底段、右肺前基底段、右肺外基底段、右肺后基底段、左肺尖后段、左肺前段、左肺上舌段、左肺下舌段、左肺背段、左肺内前基底段、左肺外基底段、左肺后基底段。3.算法处理：基于标注结果，根据肺段标注与结节标注的重合范围，按照最大重叠范围原则，判断出结节所处肺段位置。4.数据应用：用于胸部CT影像自动分割人工智能模型的训练。

This dataset is constructed by annotating raw chest CT imaging data in DICOM format. The annotated targets include trachea, pulmonary nodules, right pulmonary artery, right pulmonary vein, left pulmonary artery, left pulmonary vein, as well as 18 pulmonary segments: right apical segment of right lung, right posterior segment of right lung, right anterior segment of right lung, right lateral segment of right lung, right medial segment of right lung, right dorsal segment of right lung, right medial basal segment of right lung, right anterior basal segment of right lung, right lateral basal segment of right lung, right posterior basal segment of right lung, left apicoposterior segment of left lung, left anterior segment of left lung, left superior lingular segment of left lung, left inferior lingular segment of left lung, left dorsal segment of left lung, left anteromedial basal segment of left lung, left lateral basal segment of left lung, left posterior basal segment of left lung. This dataset is provided for training automatic image segmentation artificial intelligence (AI) models, to enable the AI models to better identify various organs, tissues and lesions in chest CT images, extract features and discover patterns, so as to ultimately improve the accuracy, robustness and generalization ability of diagnostic AI models.
1. Data Collection: Anonymized chest CT data is obtained from medical institutions through formal cooperation agreements.
2. Data Annotation: Based on knowledge of human anatomy and medical imaging diagnostics, combined with CT imaging features, the data is annotated following the workflow of "one person annotates, two persons review, and one clinical expert confirms". All the following anatomical structures and nodules are annotated: trachea, pulmonary nodules, right pulmonary artery, right pulmonary vein, left pulmonary artery, left pulmonary vein, right apical segment of right lung, right posterior segment of right lung, right anterior segment of right lung, right lateral segment of right lung, right medial segment of right lung, right dorsal segment of right lung, right medial basal segment of right lung, right anterior basal segment of right lung, right lateral basal segment of right lung, right posterior basal segment of right lung, left apicoposterior segment of left lung, left anterior segment of left lung, left superior lingular segment of left lung, left inferior lingular segment of left lung, left dorsal segment of left lung, left anteromedial basal segment of left lung, left lateral basal segment of left lung, left posterior basal segment of left lung.
3. Algorithm Processing: Based on the annotation results, the pulmonary segment location where each nodule is located is determined according to the maximum overlap principle between the nodule annotation region and the pulmonary segment annotation regions.
4. Data Application: This dataset is used for training automatic chest CT image segmentation AI models.