摘要
针对鲁棒模糊局部信息C-均值聚类分割算法易丢失图像细节的问题,提出一种改进的核空间直觉模糊C-均值聚类算法。将像素空间邻域信息和直觉指数引入到鲁棒模糊局部信息C-均值聚类目标函数,给出改进的像素空间邻域信息约束的聚类目标函数,对其聚类目标函数最优化推导并得到新的隶属度和聚类中心迭代表达式,并设计相应的图像分割算法,以便提高图像局部信息的有效分割能力。实验结果表明,改进的核空间直觉模糊聚类分割算法相比现有鲁棒模糊局部信息C-均值聚类分割算法能获得更好的分割效果。
In view of the issues that an improve kernel space intuitionistic fuzzy C-means clustering segmentation algorithm is proposed in this paper to tackle the problem that the fuzzy c-means clustering with fuzzy weighted factor and kernel metric （KWFLICM） segmentation algorithm could not keep the image details well. By introducing the pixel spatial neighbor information and the hesitation degree into the objective function of the KWFLICM algorithm, a new pixel spatial neighbor information constraints clustering objective function is proposed. By optimizing the clustering objective function, the new clustering center and membership iterative expressions are obtained. The corresponding image segmentation algorithm is presented to improve the ability of effective segmentation of image local information. Experimental results demonstrate that it can get the better results than that of KWFLICM.
出处
《西安邮电大学学报》
2015年第6期45-50,共6页
Journal of Xi’an University of Posts and Telecommunications
基金
国家自然科学基金重点资助项目(61136002)
陕西省自然科学基金资助项目(2014JM8331
2014JQ5138)
关键词
模糊C-均值聚类
像素空间邻域信息
核空间
直觉模糊集
fuzzy C-means cluster, pixel spatial neighbor information, kernel space, intuitionistic fuzzy set
