模糊C-均值聚类图像分割算法的一种改进

摘要

针对传统模糊C-均值聚类算法对含噪图像分割时未充分考虑空间信息的问题，提出一种改进的模糊C-均值聚类算法，将图像的局部和非局部两种空间信息引入到模糊C-均值聚类算法的目标函数中，以使两种空间信息在含噪图像分割中发挥互补作用。将改进算法应用于不同含噪图像的分割实验，结果表明图像像素的均方误差均比改进前有所降低。%In view of the problem that traditional fuzzy c-means (FCM)clustering segmentation algorithm does not consider the spatial information of noisy image sufficiently,an improved fuzzy c-means (FCM)clustering segmentation algorithm is proposed in this paper.The improved algorithm introduces both local and non-local spatial information into the objective function,and the two spatial information can then play a positive and complementary role in guiding noisy image segmentation.The improved algorithm cab be successfully used for different noisy image segmentation,and segmentation results show that the mean squared error of image pixels are greatly reduced.