医学图像自动分割的神经网络算法

提出了一种基于Ｋｏｈｏｎｅｎ聚类神经网络的图像分割算法。首先论述了Ｋｏｈｏｎｅｎ聚类神经网络的基本原理，在此基础上对其进行了改进，将其用于医学图像分割中。针对聚类中心初始值选取的盲目性，提出了初始值优选法，大幅度提高了分割算法的速度。实验表明，本文提出的算法能快速、准确地完成医学图像的自动分割。

Neural Network Method for Medical Image Segmentation

Having considered the clustering neural network, by applying it to medical image segmentation, an amended clustering neural network algorithm is presented. The neural network method for medical image segmentation is also established for the first time. By using the new method, the experiment proves that the medical picture can be segregated quickly and correctly.