脑灰质的一种自动分割算法研究

A Fast Segmentation Algorithm for Gray Matter

针对MRI脑部图像,提出一种脑灰质的自动分割算法。根据观察和研究发现,人脑MRI中的脑灰质部分存在一种分布形式,把这种分布形式抽象成一种曲线形式从而进行分割,可以得到一种较为准确的脑部图像分割效果。自动分割算法主要分为3个步骤:首先,计算出图像的直方图并对直方图进行模糊;其次,利用模糊聚类算法计算出聚类中心;最后,根据模糊聚类中心计算出脑灰质区域的曲线,从而实现阈值分割。自动分割算法经过大量模拟和真实MRI脑部图像进行了验证。

An automatic segmentation method for gray matter of human brain MRI is proposed. According to the observation and research, it is found that there is a kind of distribution in the gray matter of human brain MRI, which can be abstracted into a kind of curve to obtain a more accurate brain image segmentation. The algorithm is mainly divided into three steps： first, to calculate the image histogram and the fuzzy histogram; secondly, using fuzzy clustering algorithm to calculating the clustering center; finally, according to the clustering center to calculate the region of cerebral gray matter curve to realize threshold segmentation. The proposed algorithm is validated by extensive simulations and real human brain MRI.