中文摘要：

为了克服传统模糊C-均值（fuzzy C-means，FCM）聚类算法特征描述单一、易受复杂灰度影响而出现误分割的缺点，将万有引力和图像局部熵融入到FCM算法。算法首先引入图像局部信息熵来描述节点（像素点）间的特征，同时计算新节点的同质值；其次，将该同质值看做新节点的质量，节点之间通过万有引力算子形成关联，使节点灰度特征和节点空间位置特征有效结合，以此解决传统FCM算法节点特征描述孤立的缺陷。最后，对三类典型的灰度分布不均的医学图像进行仿真实验，结果表明改进算法获得了更加精确的分割结果。

英文摘要：

To overcome the shortcomings of the traditional fuzzy C-means （FCM） clustering algorithm which were simple image feature description and easy distributed by complex grey influence with wrong segmentation, this paper proposed an improved FCM algorithm for image segmentation, combined with universal gravitation principle and local entropy theorem. Firstly, it introduced the image local entropy to accurately measure image node property between two adjacent nodes, and meanwhile computed the node homogeneous value. Then it was taken as the node quality, formed closely relationship using gravity algorithm which made the node grey feature and spatial position combine effectively. The above method solved the problem of the description of node feature isolation of traditional FCM algorithm. Finally, the simulation results show that the presented algorithm can obtain more precise segmentation results from three types of grey non-uniform distribution medical images.