基于厚度特征模糊C均值算法的膀胱肿瘤疑似区域三维分割方法研究  

Fuzzy C-means algorithm using wall thickness feature for 3D bladder tumor suspected region segmentation

作　　者：肖丹[1] 刘洋[1] 张国鹏[1] 张曦[1] 卢虹冰[1] 

机构地区：[1]第四军医大学生物医学工程学院,西安710032

出　　处：《医疗卫生装备》2015年第3期1-4,31,共5页Chinese Medical Equipment Journal

基　　金：国家自然科学基金资助项目(81230035;81071220)

摘　　要：目的:使用模糊C均值算法对膀胱壁厚度特征进行聚类,有效地分割出三维膀胱肿瘤疑似区域。方法:获取膀胱MRI数据,应用Level Set分割算法对膀胱内、外壁进行分割,然后计算膀胱壁的三维厚度,通过对膀胱壁厚度聚类得到三维肿瘤疑似区域。结果:由该方法分割得到的疑似区域比较准确,与临床放射医师勾画的肿瘤区域有90%以上的重叠率。结论:该方法可有效地分割出三维膀胱肿瘤疑似区域,具有一定的有效性,但尚需要采集更多的数据来进行验证。Objective To use Fuzzy C-means（FCM） method based on wall thickness feature to segment the 3D suspected regions. Methods With bladder MRI data acquired, Level Set segmentation algorithm was applied to cutting inner and outer surfaces of the bladder wall in MRI T23 D sequence. The bladder wall thickness was calculated by 3D thickness measurement algorithm. Suspected regions were divided by the proposed FCM method, and a method using the threshold values to resolve the clustering number in FCM was proposed. Results The suspected regions by the method had a rate of 90% to overlap those determined by radiologists. Conclusion The method can effectively make clear 3D bladder tumor suspected region, while need data for further validation.

关 键 词：三维分割 模糊C均值算法 膀胱肿瘤 膀胱壁厚度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R737.14[自动化与计算机技术—计算机科学与技术]