合理利用空间信息的MRI脑部图像分割

在合理利用空间信息的基础上,提出了一种更准确,紧致性和分离性更好的分割算法。该算法首先定义一个空间函数,并在其中引入一个控制参数,该参数可以对噪声点、边缘点以及区域内部的点进行区别对待,然后用空间信息更新隶属度。实验结果表明,该算法效果要明显优于sFCMpq算法及其改进算法(EsFCMpq)。

The wise use of spatial information for MRI brain segmentation

Based on the wise use of spatial information, this paper proposes a more accurate segmentation algorithm with samples compact within one cluster and separated among different clusters. The algorithm firstly defines a spatial function by introducing a control parameter, which can make the noise points, edge points and points within the region be treated differently. Then it uses the spatial information to update the membership function. Experimental results show that the proposed algorithm outperforms the sFCMpq algorithms and its improved algorithm EsFCmpq in various cluster validity functions.