基于灰度重构和PSO的FCM分割MRI  

FCM segmentation MRI based on grayscale reconstruction and PSO

作　　者：周同驰 李征 瞿博阳[1] 李召 宋晓炜[1] ZHOU Tong-chi;LI Zheng;QU Bo-yang;LI Zhao;SONG Xiao-wei(School of Electronic and Information,Zhongyuan University of Technology,Zhengzhou 450007,China)

机构地区：[1]中原工学院电子信息学院,河南郑州450007

出　　处：《计算机工程与设计》2021年第9期2568-2576,共9页Computer Engineering and Design

基　　金：国家自然科学基金项目(61673404、61876169);河南省高等学校重点科研基金项目(19A120014);河南省高校创新人才基金项目(16HASTIT033);河南省高等学校青年骨干教师培养计划基金项目(2016GGJS-094);中国纺织工业联合会科技指导性基金项目(2017054、2018104、2018107);河南省高校科技创新团队支持计划基金项目(18IRTSTHN013)。

摘　　要：为解决模糊C均值(FCM)模型对噪声敏感和随机初始聚类中心影响分割结果的问题,提出一种基于灰度重构和自适应粒子群优化的FCM算法。使用图像中每两个像素的空间和灰度信息,构造一种加权方式对灰度进行加权和,进行Top-hat和Bottom-hat变换,得到灰度重构图像;对粒子群算法的惯性权重进行改进,使其根据迭代情况自适应更新粒子速度,避免陷入局部最优;将改进的粒子群算法与FCM相融合,粒子位置作为聚类中心,对灰度重构后的图像进行分割。以磁共振图像为例进行实验,并与同类型算法进行对比,所提算法取得了较好的分割结果。To solve the problem that the fuzzy C-means(FCM)model is sensitive to noise and the random initial clustering center affects the segmentation results,an FCM algorithm based on gray scale reconstruction and self adaptive particle swarm optimization(SAPSO)was proposed.The spatial and grayscale information of every two pixels in the image was used to construct a new weighted method to carry out the weighted sum of grayscale,and top-hat and bottom-hat transformation was carried out to obtain the grayscale reconstruction image.The inertia weight of the particle swarm optimization algorithm was improved,so that it adaptively updated the particle speed according to the iteration situation and avoided falling into the local optimum.The improved particle swarm optimization algorithm was fused with FCM,and the particle position was used as the clustering center to segment the grayscale reconstructed image.Taking magnetic resonance image(MRI)as an example,and comparing with similar algorithms,the proposed algorithm achieves good segmentation results.

关 键 词：模糊C均值 灰度重构 粒子群优化 图像分割 磁共振图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]