基于改进的免疫模糊聚类方法的医学图像分割  

An improved fuzzy clustering algorithm based on immune in medical image segmentation

作　　者：吴宇翔[1] 龚涛[1] 梁文宇[1] 

机构地区：[1]东华大学信息科学与技术学院,上海201600

出　　处：《微型机与应用》2016年第6期51-53,57,共4页Microcomputer & Its Applications

摘　　要：传统模糊C-均值聚类算法需要输入初始聚类中心,但是输入错误的初始聚类中心会产生较差的图像分割结果。对此提出一种改进的医学图像分割算法——基于免疫模糊聚类的医学图像分割。该算法能够快速有效地找出合适的初始聚类中心值,使之最大可能地趋近于理想值,从而大大提高算法的效率,避免陷入局部解。同时,将免疫克隆选择算法融入到模糊聚类算法中。实验结果表明,该算法能快速有效地找到合适的初始聚类中心,能有效提高搜索效率和准确率,得到较理想的分割效果。It is necessary to enter the initial center value before use the traditional fuzzy C-means clustering（FCM）,but error initial center value will bring error image segmentation. In this paper,an improved fuzzy clustering algorithm based on immunein medical image segmentation has been presented. It can find the suitable initial center value quickly to improve the efficiency and accuracy and avoid trapping in local optimum.Besides,this algorithm adds immune clone selection in it. The simulation results show that the centers of clusters are adaptively decided by using this improved algorithm. It makes perfect image segmentation result and improves the efficiency and accuracy.

关 键 词：免疫模糊聚类 图像分割 克隆选择 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]