摘要
模糊C-均值聚类算法是一种无监督图像分割技术,但存在着初始隶属度矩阵随机选取的影响,可能收敛到局部最优解的缺点。提出了一种粒子群优化与模糊C-均值聚类相结合的图像分割算法,根据粒子群优化算法强大的全局搜索能力,有效地避免了传统的FCM对随机初始值的敏感,容易陷入局部最优的缺点。实验表明,该算法加快了收敛速度,提高了图像的分割精度。
The Fuzzy C-means （FCM）clustering algorithm is a no-supervise image segmentation algorithm.But it is sensitive to initial clustering membership subordination matrix and likely converges into the local minimum.A new image segmentation algorithm is proposed,which combines the particle swarm optimization （PSO）and FCM clustering. A powerful global search capabilities based on PSO algorithm ,it effectively avoid the traditional FCM sensitive to rand initial values, vulnerable to the shortcomings of local optimization. It is shown from the experiments that our proposed algorithm accelerate the speed of convergence, improve the quality of image segmentation.
出处
《电子设计工程》
2012年第18期167-169,共3页
Electronic Design Engineering
基金
陕西省工业攻关项目(2011K06-35)
关键词
粒子群优化
模糊C均值聚类
全局搜索
图像分割
particle swarm optimization
fuzzy C-mean clustering
global search
image segmentation
