基于核模糊聚类优化算法的脑核磁共振图像分割研究  

Research of MR Brain Image Segmentation Based on Kernel Fuzzy C-means Optimization Algorithm

作　　者：万春圆 叶明全 姚传文[1] 徐争元 WAN Chun-yuan;YE Mingquan;YAO Chuan-wen(School of Medical Information,Wannan Medical College,Wuhu 241002,Anhui Province,P.R.C.)

机构地区：[1]皖南医学院医学信息学院,241002 [2]皖南医学院健康大数据挖掘与应用研究中心,241002 [3]皖南医学院医学影像学院,241002

出　　处：《中国数字医学》2020年第11期10-15,共6页China Digital Medicine

基　　金：国家自然科学基金面上项目(编号:61672386);安徽省自然科学基金面上项目(编号:1708085MF142);教育部人文社会科学研究规划基金(编号:16YJAZH071)。

摘　　要：目的:脑核磁共振(MR)图像中普遍存在噪声的影响,传统的核模糊C-均值(Kernel Fuzzy C-means,KFCM)算法无法得到理想的脑组织分割结果,为此提出一种基于核模糊聚类优化算法的分割模型。方法:首先通过粒子群算法确定KFCM的初始聚类中心,然后利用自适应中值滤波消除图像中的噪声,最后采用该模型分别对不同的图像进行实验。结果:该方法不仅能迅速确定图像的初始聚类中心,并且有效地消除图像中的噪声。结论:与传统KFCM算法相比,提出的模型具有更高的精确度和分割效率。Objective:The traditional Kernel fuzzy C-means(KFCM)clustering method cannot obtain the ideal brain segmentation results in the brain magnetic resonance(MR)image.In this paper,a new method of Kernel Fuzzy C-means based on optimization algorithm for MR brain image segmentation is proposed.Methods:The method through the Particle Swarm Optimization to determine the clustering center to improve the KFCM clustering effect,and uses adaptive median filter to remove the noise.By this method,20 cases of brain MR images provided by brain web were segmented.Results:This method can quickly determine the initial clustering center of the image and effectively eliminate the noise.Conclusion:Experimental results show that compared with the traditional KFCM image segmentation method,this method has better segmentation accuracy and segmentation efficiency.

关 键 词：脑磁共振图像 核模糊C-均值 图像分割 自适应中值滤波 粒子群算法 

分 类 号：R319[医药卫生—基础医学]