首页 | 本学科首页   官方微博 | 高级检索  
     

改进的基于遗传模糊C均值聚类的图像分割算法
引用本文:仲崇峰,刘智. 改进的基于遗传模糊C均值聚类的图像分割算法[J]. 长春理工大学学报(自然科学版), 2014, 0(2): 62-67
作者姓名:仲崇峰  刘智
作者单位:长春理工大学电子信息工程学院,长春130022
基金项目:吉林省教育厅科技基金资助项目(吉教科合字[2012]第21号)
摘    要:模糊C均值(Fuzzy C-Means,FCM)聚类算法已广泛应用于图像分割领域,其本质是一种局部搜索算法,采用迭代爬山算法寻找最优解,对初始聚类中心敏感,很容易陷入局部极优值,且没有考虑图像的空间邻域信息,对噪声敏感。本文提出了改进的基于遗传模糊聚类的图像分割算法,利用遗传算法的全局寻优能力来克服FCM算法容易陷入局部极优值问题;并在FCM算法的目标函数中添加空间邻域信息来约束隶属度函数从而提高对噪声的鲁棒性,使分割更加符合期望。实验结果表明本文算法的有效性,图像分割时具有较强的抗噪能力和较好的分割效果。

关 键 词:模糊C均值  遗传算法  空间邻域信息  图像分割  噪声

An Improved Image Segmentation Algorithm Based on Genetic Fuzzy C-Means Clustering
ZHONG Chongfeng,LIU Zhi. An Improved Image Segmentation Algorithm Based on Genetic Fuzzy C-Means Clustering[J]. Journal of Changchun University of Science and Technology, 2014, 0(2): 62-67
Authors:ZHONG Chongfeng  LIU Zhi
Affiliation:(School of Electronics and Information Engineering, Changchun University of Science and Technolody, Changchun 130022)
Abstract:Fuzzy C-Means (FCM) clustering algorithm has been widely used in the field of image segmentation, be-cause its nature is a local search algorithm. As using iterative climbing to find the optimal solution,so it is sensitive to initial cluster center and easy to fall into local excellent value. Without considering the image spatial information, it is sensitive to noise. Thus,an improved image segmentation algorithm based on genetic Fuzzy C-Means clustering is pro-posed, using genetic algorithm’s global optimization ability to overcome the problem of falling into local optimal value. It improve the robustness of image noise through adding the neighborhood spatial information to FCM objective function to constrain the membership function,so that the results of segmentation are more in line with expectation. Experimen-tal results indicate that the effectiveness of this algorithm can have a strong anti-noise ability and get better segmenta-tion effectiveness.
Keywords:Fuzzy C-Means  genetic algorithm  spatial neighboring information  image segmentation  noise
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号