结合模糊C均值聚类与图割的图像分割方法

本文针对模糊C均值聚类没有考虑像素空间信息的不足，提出一种结合模糊C均值聚类与图割的图像分割方法。本文以图割理论为基础，考虑到像素的空间信息，建立一个关于标号的全局能量函数，以FCM聚类中心为终端建立多终端网络图，该网络通过 扩展移动算法求解全局最小或近似最小能量函数所对应的标号函数 ，在各类间重新划分所有像素点，实现目标正确分割。实验表明，本文方法在分割精度、性能、抗噪性等方面均有较大改进。

Image segmentation approach of combining fuzzy clustering and graph cuts

This paper proposed an image segmentation method combining fuzzy C means clustering and graph cuts. Based on graph cuts, considering the pixel space, establish a global energy function concerning labels, and make the FCM clustering center as terminal to found a multi-terminal network, get the global minimum of the energy function or the approximate minimum of the label function by the α expansion moving algorithm .Re demarcate all pixels between all classes and get the right targets. Experimental results show that the method has great improvement in accuracy, performance and noise immunity.