基于直方图统计快速FCM算法的医学脑部图像分割

针对模糊聚类算法（FCM）在脑部MRl图像分割中存在计算量大的缺点，提出了一种结合直方图峰值信息和统计信息的快速FCM（HF—KFCM）算法。算法首先利用多尺度窗口遍历的方法找到直方图的峰值点，然后将其作为模糊聚类的初始化中心。并采用基于直方图统计的快速聚类方法减少每一次迭代的运算量。仿真结果表明，相比于FCM算法和其他改进FCM算法，该算法的分割结果在聚类有效性和模糊性上提高显著。

Medical Brain Images Segmentation by Histogram Statistical Fast FCM Algorithm

For the shortcomings of huge calculation in the brain MRI images segmentation with Fuzzy C-Mean algorithm （FCM）, a new algorithm combined with histogram peak and statistical information of Fast Fuzzy C-Mean algorithm （HF- KFCM） is introduced. Firstly, the method of multi-scale window traverse is used by this algorithm to find the histogram peaks as the fuzzy clustering initialization centre. Then, the fast FCM method based on histogram statistic is used to reduce each iteration calculation. The simulation results show that compared with FCM algorithm and other improved FCM algorithms, the proposed algorithm can improve significantly in the effectiveness of fuzzy and clustering.