基于FCM聚类算法的快速水平集图像分割仿真  

Fast Level Set Image Segmentation Simulation Based on FCM Clustering Algorithm

作　　者：刘铭 孙守云 LIU Ming;SUN Shou-yun(Software College,Sichuan University,Chengdu Sichuan 610044,China;Institute of Software,Chinese Academy of Sciences,Beijing 100080,China)

机构地区：[1]四川大学软件学院,四川成都610044 [2]中国科学院软件研究所,北京100080

出　　处：《计算机仿真》2019年第11期378-382,共5页Computer Simulation

摘　　要：为解决当前水平集图像分割方法中由于噪声过高引起的聚类和分割效果差的问题,提出基于FCM聚类算法的快速水平集图像分割方法。在水平集图像分割中引入聚类能量和正规化能量概念,结合小波变换方法抑制图像中的噪声;利用FCM聚类方法设计图像分割模型,利用模型进行模糊聚类以及交替曲线平滑过程,获取能量泛函的极小值,以达到快速水平集图像分割的目的;改进模型中边界截止函数,并通过灰度值与隶属度之间的关系修正已分割图像的边缘信息,提高分割质量。实验结果表明,所提方法能够有效降低噪声对分割结果所产生的影响,获取理想的图像分割结果。In order to solve the problem of poor clustering effect and segmentation effect caused by excessive noise in current methods, a method to fast segment the level-set image based on FCM clustering algorithm is proposed. In the level-set image segmentation, the concepts of clustering energy and regularization energy were combined with the wavelet transform method to suppress the noise in image. Then, FCM clustering method was used to design the image segmentation model, which was used to perform fuzzy clustering and alternate the curve smoothing process, so as to get the energy functional minimum value. After that, the fast horizontal set image segmentation could be achieved. Moreover, the boundary function in model was improved, and the edge information of the segmented image was corrected by the relationship between the gray value and the membership degree. Thus, the segmentation quality was improved. Simulation results prove that the proposed method can effectively reduce the influence of noise on the segmentation results and obtain the ideal result of image segmentation.

关 键 词：快速水平集 图像分割 噪声抑制 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]