基于改进FCM的医学图像分割

Medical Image Segmentation Based on Modified FCM Algorithm

为解决模糊C均值聚类(FCM)算法在图像分割尤其是医学图像分割中存在的计算量大、运行时间过长的问题,提出了一种改进方法。通过数据约减,即通过对相近的像素进行量化并聚合来减少像素个数,从而降低运算量。该方法用于人脑磁共振图像的分割比传统FCM算法的运算速度提高了50￣100多倍,并且选择合适大小的量化箱不会影响算法的分割效果。

The fuzzy C means clustering （FCM） algorithm requires a long time to segment images, especially medical images, due to processing the large dataset. This paper discusses a modified algorithm based on data reduction, which is able to reduce the number of pixels by aggregating similar ones. The reduction in the amount of clustering data allows a partition of the data to be produced faster. The algorithm is applied to the problem of segmenting brain magnetic resonance images into different tissue types. Average speed-ups of as much as 50-100 times a traditional implementation of fuzzy c-means were obtained, while producing partitions that are equivalent to those produced by fuzzy c-means.