孤立性肺结节计算机辅助诊断系统的研究  

Research on Computer-aided Diagnosis System of Solitary Pulmonary Nodules

作　　者：朱燕杰[1] 谭永强[1] 张建国[1] 

机构地区：[1]中国科学院上海技术物理研究所医学影像实验室,上海市玉田路500号200083

出　　处：《中国数字医学》2010年第7期32-35,共4页China Digital Medicine

基　　金：国家自然科学基金(编号:30570512)~~

摘　　要：孤立性肺结节是肺部的常见病灶,其良恶性诊断具有重要医学意义。利用计算机辅助诊断系统可提高医生诊断的准确率,减少漏诊率。构建了肺结节计算机辅助诊断系统,并选取了77例已确诊的孤立性肺结节的CT图像,并对系统性能进行了研究。该辅助诊断系统提取了67个图像模式特征,通过遗传算法对特征进行选择,选出高性能的特征子集,采用支持向量机作为分类器鉴别结节良恶性。试验结果表明,该方法可获得可靠的结果。Solitary pulmonary nodules （SPNs） are common findings in thoracic imaging and the diagnosis has great significance. Using computer-aided system could improve the accuracy rate in diagnosis and reduce the false dismissal rate. In this paper, a computer-aided system based on solitary pulmonary nodules was constructed and seventy-sever biopsy-confirmed CT cases with SPNs were included to test the performance of the system, A total of sixty-seven features were extracted in this study and high performance subsets of features were selected by genetic algorithm. The classifier of Support Vector Machine （SVM） was constructed with selected features to distinguish malignant from benign nodules. Experimental results verified the effectiveness of the presented system.

关 键 词：特征提取 支持向量机 孤立性肺结节 计算机辅助诊断 

分 类 号：R979.1[医药卫生—药品]