| 基于自适应初始值的FCM聚类图像分割 |
| |
| 引用本文： | 王兴,冯子亮. 基于自适应初始值的FCM聚类图像分割[J]. 微机发展, 2010, 0(3): 101-104 |
| |
| 作者姓名： | 王兴 冯子亮 |
| |
| 作者单位： | 四川大学图像图形研究所; |
| |
| 基金项目： | 国家自然科学基金重点资助项目(60736046) |
| |
| 摘 要： | FCM聚类算法对初始值敏感，不良的初始值会导致算法的收敛速度过慢和收敛到局部极值。将FEM算法用于图像分割处理时，初始值的选择是一个难点。文中提出了一种使用自适应初始值的FCM聚类图像分割算法，该方法利用图像的直方图特性建立候选聚类中心集，通过初始化准则函数检验候选集得到合适的聚类中心和聚类数目，并根据最大隶属度原则分割图像，得到了较好的分割效果。理论分析和实验表明文中方法收敛速度快，分割准确，自适应性很强。
|
| 关 键 词： | 图像分割 FCM聚类 自适应性 |
An Image Segmentation Algorithm Based on Adaptive Initialization FCM Clustering |
| |
| WANG Xing,FENG Zi-liang. An Image Segmentation Algorithm Based on Adaptive Initialization FCM Clustering[J]. Microcomputer Development, 2010, 0(3): 101-104 |
| |
| Authors: | WANG Xing FENG Zi-liang |
| |
| Affiliation: | WANG Xing,FENG Zi-liang(Institute of Image , Graphics,Sichuan University,Chengdu 610064,China) |
| |
| Abstract: | A method of image segmentation based on improved FCM clustering is presented in this paper,which is characterized by its good partition peculiarity,low computing cost and strong self-adapting.Unlike many existing clustering algorithms,this method does not require the number of cluster-centers before clustering,and it can use the image's information to find all cluster-centers automatically by two steps.First,using image's histogram generates a set of possible cluster-center;then,decreases every unbefitting ... |
| |
| Keywords: | image segmentation FCM clustering self-adapting |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
