摘要： 运用计算机技术辅助诊断脑中风出血病人对于精确计算出血量、及时挽救生命有着重要的意义。有时由于出血点小、边缘模糊,给 图像的分割造成一定的困难。根据图像的特征,该文运用基于Gustafson-Kessel模糊C均值聚类(FCM)两步分割算法较为准确地对出血区域 及水肿带进行了分割,同时,在分割开始时运用了TT图谱配准,从而减少了分割影响因素,提高了分割的有效性。

关键词: 脑出血CT图像, Gustafson-Kessel模糊C均值聚类, 自动分割

Abstract: An automatic segmentation method of CT image of spontaneous intracerebral brain hemorrhage is introduced in this paper. The algorithm could be divided into two levels. Before segmentation, the dataset is registered with the TT atlas to reduce the affected facts. After that, Gustafson-Kessel fuzzy C-means is used twice in two levels to do the segmentation. First the Global segmentation, then the local. The method can be mostly and automatically except selection of hemorrhage and edema region, and can improve the accuracy and efficiency of the segmentation.

中图分类号: 

• TP391.41