针对传统模糊C-均值聚类算法对复杂的医学、遥感图像难以获得满意分割效果问题,将图像模糊C-均值聚类引入图像分割问题研究中,提出了基于直方图的图像模糊聚类快速分割算法。将越南学者Le提出的分布式图像模糊聚类算法目标函数进行简化,得到图像模糊聚类算法目标函数;采用拉格朗日乘予法获取其迭代求解所对应的隶属度、中立度、拒分度和聚类中心表达式,设计图像模糊聚类算法并对其收敛性进行了证明。通过复杂医学和遥感图像的分割测试结果表明,新的分割算法相比现有的模糊C-均值聚类分割算法和直觉模糊C-均值聚类分割算法具有更好的分割性能。
A fast segmentation algorithm based on histogram of picture fuzzy clustering is'proposed by introducing the picture fuzzy C- means clustering into the image segmentation since it is difficult for the traditional fuzzy C- means clustering algorithm to obtain satisfactory segmentation results for complex medical and remote sensing images. Firstly, the objective function of distributed picture fuzzy clustering algorithm proposed by Le is simplified to get the objective function of picture fuzzy clustering algorithm. The corresponding positive degree, neutral degree, refusal degree and cluster center expression are obtained by the Lagrange multiplier method. The fuzzy picture clustering algorithm is designed and its convergence is proved. The segmentation results of complex medical and remote sensing images show that the new segmentation algorithm has better segmentation performance than the existing fuzzy C-means clustering algorithm and intuitionistic fuzzy C-means clustering algorithm.