摘要
针对传统的FCM算法随机获取初始聚类中心与分类类别数的缺陷问题,提出了一种获取初始聚类中心与分类类别数的方法,并采用交叉熵测度准则进行FCM聚类,对彩色图像进行分割,提取有意义区域。实验结果表明,该方法不仅能够提高算法的聚类速度与算法的普适度,而且可以改善图像的聚类效果。与传统的FCM算法相比,该算法更易于实现彩色图像有意义区与背景的分离,分割效果令人满意。
Considering the shortcomings of traditional FCM algorithm, a new method is proposed to select initial clustering center and the number of clusters. Besides that, the theoryof cross-entropy distance is introduced to operate procedure of FCM algorithm to segment color images. Experimental results show that, the new algorithm not only improves the initial clustering center, but also enhances the segmentation precision and the universality principle. Comparing with the traditional FCM algorithm, the new algorithm can easily realize the separation of background and meaningful region, and the results are satisfied.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第18期4082-4084,4104,共4页
Computer Engineering and Design
基金
江苏省高校自然科学基金项目(06kja14003)
关键词
模糊C均值聚类算法
彩色图像分割
交叉熵距离
特征散度距离
隶属度准则
fuzzy C-means clustering algorithm
color image segmentation
cross-entropy distance
feature divergence distance
degree of membership rule
