摘要
在图像分割中,针对FCM算法存在聚类数目需要预先给定、收敛速度慢等缺点,本文把快速模糊C均值聚类算法和随机游走算法相结合,具体方法为先采用快速模糊C均值聚类算法对图像进行预分割,以便获得聚类中心的位置,然后将该中心作为随机游走的种子点,再进行图像分割,实验结果得到了较为满意的预期效果,证明该方法是可行的.本文的研究为快速FCM实现自适应性和开发图形图像预处理系统提供了技术支持与理论依据.
As far as image segmentation, the defeat of the number of clusters for FCM algorithm is reeded to be improued. In this paper, the fast fuzzy C-means clustering and random walk algorithm are combined to solve the problem of image segmentation. Firstly, the fast FCM for image pre-segmentation to obtain the number of clusters and cluster central location as the seed points of random walk firstly. Then, for image segmentation,experimental results show that this method is feasible, and get a more satisfactory desired purpose. Results of this study achieve self-adaptive and fast FCM develop graphical image preprocessing system provides technical support and theoretical basis.
出处
《广东技术师范学院学报》
2016年第2期48-52,共5页
Journal of Guangdong Polytechnic Normal University
基金
广州市科技和信息化局2013年应用基础研究专项(2013J4100073)
