项目名称: 基于混合成像的孤立性肺结节计算机辅助诊断方法
项目编号: No.61202163
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 计算机科学学科
项目作者: 赵涓涓
作者单位: 太原理工大学
项目金额: 24万元
中文摘要: 目前,肺癌的死亡率非常高,主要原因是肺癌早期诊断的漏诊率和误诊率居高不下。因为肺癌的早期大多表现为孤立性肺结节,而孤立性肺结节中有大约20%-40%为恶性。如果能在早期发现并确诊孤立性肺结节的良恶性,肺癌的治愈率将会有效提高。孤立性肺结节的早期诊断需要医师从海量的肺部功能影像(单光子发射断层成像)和结构影像(病灶结节结构造影成像)的医学影像数据中,结合专业的诊断经验进行综合判断。本课题通过研究量化分析孤立性肺结节功能影像和结构影像的混合成像图像视觉特征;利用机器学习方法使图像视觉特征、影像学属性特征和医学诊断语义有效融合,建立结节特征诊断关联模型;并进一步建立计算机辅助诊断模式库;自动或半自动地实现孤立性肺结节的早期诊断;有效提高孤立性肺结节的诊断效率和准确率,最终为肺癌的早期自动诊断提供量化和可视化的技术支持。项目对肺癌早期诊断具有一定的理论意义和学术价值。
中文关键词: PET-CT混合成像;孤立性肺结节;计算机辅助诊断模型;;
英文摘要: Nowadays, the lung cancer mortality rate are astonishing high. One main reason is due to the high misdiagnosis rate of early stage lung cancer. Most of the lung cancer at early stage is manifested as a solitary pulmonary nodule, and about 20%-40% of the solitary pulmonary are malignant. If we can find and diagnose whether the solitary pulmonary nodules are benign or malignant for early detection, the cure rate of lung cancer will be effectively improved. In the early diagnosis of solitary pulmonary nodule, physicians should make a comprehensive judgment from massive pulmonary functional image (single photon emission tomography) data and structural image (lesions nodular structure imaging) data using their professional diagnostic experience. This project can establish nodule diagnosis features related model by researching and analyzing the hybrid imaging image visual features of functional and structural images of Solitary Pulmonary Nodules(SPN) and using the machine learning methods to integrate quantitative information of functional and structural imaging features of the lungs and medical diagnostic information. Furthermore, this project will develop computer-aided diagnostic Pattern library, will achieve automatic or semi-automatic diagnosis of SPN; will effectively improve the efficiency and accuracy of diagn
英文关键词: PET-CT Hybrid Imaging;Solitary Pulmonary Nodules;computer-aided diagnosis model;;