论文部分内容阅读
根据肺结节的特征模型对感兴趣区域(ROI)进行分类来检测肺结节。首先,使用蚁群算法构造初步的规则集合,根据条件元素的相关度对规则进行修改确定最终的规则集合。然后,对规则集合中的特征进行建模,定义特征变量表示选定特征。最后,根据ROI的这些特征变量值以及总体特性,利用蚁群算法对ROI进行分类。分类中根据蚂蚁尸体堆积模型确定分类中心,采用蚂蚁觅食模型进行分类。仿真结果表明,该算法检测肺结节具有较高的敏感性,对医生后续的肺癌诊断有指导意义。
Pulmonary nodules were detected by classification of regions of interest (ROI) based on the feature models of pulmonary nodules. First, the initial rule set is constructed using the ant colony algorithm, and the rule is modified according to the relevancy of the condition elements to determine the final rule set. Then, the features in the rule set are modeled and the feature variables are defined to represent the selected features. Finally, the ROI is classified using ant colony algorithm based on these characteristic variable values of the ROI and its overall characteristics. According to the classification of ants dead body accumulation model to determine the classification center, the use of ants foraging model for classification. The simulation results show that this algorithm has a high sensitivity for detection of pulmonary nodules and is of guiding significance for the follow-up diagnosis of lung cancer.