FCM和Level Set在医学影像分割中的应用

摘要

目的 探讨RGAC-M(改进的区域几何活动轮廓模型),用于更准确地分割医学图像.方法 通过分析区域几何活动轮廓模型(RGAC)在医学图像分割中存在的缺陷,并对其区域项进行改进,提出了RGAC-M模型.该模型采用多种子初始化方式,降低了算法对初始条件的敏感性,同时也减少了人工干预.结果 利用RGAC-M对多种脑组织(灰质、白质、脑脊液以及背景等)进行分割,取得了良好效果.结论 RGAC-M可减少图像分割的迭代次数,提高图像分割速度和图像质量.%Objective To explore RGAC-M (Region Based Geometric Active Contour-Model), in order to divide medical images more accurately. Methods By analyzing the existing defects of RGAC in medical image segmentation and improving its areas, RGAC-M is put forward, namely, by adopting many ways of substate initialization, reduces the sensibility of arithmetic to initial conditions, and meanwhile decreases manual intervention. Results In the way of various of brain (cinerea, alba, cerebrospinal fiuid, background and so on) segmentation, RGAC-M got good result. Conclusion RGAC-M could reduce iterations of image segmentation and improve segmentation speed and image quality.