结合区域信息的双抑制FCM聚类图像分割  

Double suppressed fuzzy C-means clustering image segmentation combined with regional information

作　　者：兰蓉 胡天隆 赵强 LAN Rong;HU Tian-long;ZHAO Qiang(School of Communications and Information Engineering,Xi’an University of Posts and Telecommunications,Xi’an 710121,China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121

出　　处：《计算机工程与设计》2022年第6期1740-1748,共9页Computer Engineering and Design

基　　金：国家自然科学基金项目(61571361、61671377);西安邮电大学西邮新星团队计划基金项目(xyt2016-01)。

摘　　要：针对抑制式模糊C均值聚类算法在进行图像分割时出现的收敛性能较差和像素错误分割问题,提出一种结合区域信息的双抑制模糊C均值聚类图像分割算法。对图像进行初始区域划分,针对不同的区域,提取其区域信息;利用区域信息构建修正因子,实现对模糊隶属度的初次抑制;将区域信息和数据自身的分布特性相结合,利用指数函数构建抑制因子的自适应选取公式,实现对模糊隶属度的二次抑制,进一步提高收敛性能。实验结果表明,该算法可以改善像素易错分现象,提高了收敛性能。In view of the poor convergence performance and pixel misclassification problems of suppressed fuzzy C-means clustering algorithm in image segmentation,a double suppressed fuzzy C-means clustering algorithm combined with regional information was proposed.The image was divided into initial regions,and the corresponding region information was extracted for diffe-rent regions.The region information was used to construct the correction factor and realize the first suppression of fuzzy membership.The regional information was combined with the distribution characteristics of the data itself,and an adaptive selection formula of the suppression factor was constructed using the exponential function,so as to achieve the second suppression of fuzzy membership and further improve the convergence performance.Experimental results show that the proposed algorithm can improve the misclassification of pixel,and increase the convergence performance.

关 键 词：区域信息 抑制式模糊C均值聚类 抑制因子 图像分割 模糊隶属度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]