基于K均值聚类算法提取CT图像的肺部肿瘤  

Retrieval of Lung Tumor CT Images Based on K-means Clustering Algorithm

作　　者：李玲[1] 余后强[2] 

机构地区：[1]湖北科技学院生物医学工程学院,湖北咸宁437100 [2]湖北科技学院数学与统计学院,湖北咸宁437100

出　　处：《甘肃科学学报》2015年第1期58-61,共4页Journal of Gansu Sciences

基　　金：湖北科技学院科研项目(KY13087)

摘　　要：为自动提取CT肺部肿瘤,辅助医生对患病部位进行诊断和治疗,利用K均值聚类算法自动提取肺部肿瘤和剩余肺部图像,并分别和影像医生手工分割的肺部肿瘤和剩余肺部图像作对比。结果表明,自动提取的肺部肿瘤图像与专业医学影像医生手工提取的肺部肿瘤图像在外形、灰度和方差方面非常接近,说明利用K均值聚类算法自动提取肺部肿瘤的方法是有效的、可行的。To assist doctors to diagnose and treat pulmonary diseases through the automatic retrieval of lung tumor CT images. The images of lung tumor and the remaining part were automatically retrieved by the K-means clustering algorithm and then the manually-segmented images of lung tumor and remaining part were contrasted. The images of lung tumor automatically-retrieved and the one manually-segmented by the doctor were compared and contrasted in shape, gray level and variance, and the results were very close. The method of automatically retrieving lung tumor images by the K-means clustering algorithm was effective and feasible.

关 键 词：CT图像 K均值聚类算法 肺部肿瘤 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] R563.2[自动化与计算机技术—控制科学与工程]