基于聚类神经网络算法的医学图像分割

Medical image division based on clustering nerve network algorithm

图像分割是医学图像处理中的一个难题,它是指按照一定的原则将一幅图像或景物分成若干个部分或子集的过程.它是医学图像处理中极为重要的内容之一,是实现图像测量、配准、融合以及三维重建的基础,在临床诊断中也起着越来越重要的作用.使用 VISUAL C++实现了 Kohonen 聚类神经网络算法,并对此算法进行了改进优化,在此基础上设计了单窗口模式的"Division"图像处理应用程序进行图像分割,达到了良好的效果,有助于医生的临床诊断.

Image division is in great importance in medical image processing,which is applied to divide a image or scenery into several parts or subset based on certain fixed principle.It is the basis of achieving image measurement, registration,fusion and 3-D reconstruction and has obtained much more attention in clinical diagnosis.Based on VISUAL C++ to develop the Kohonen clustering nerve network arithmetic,single-window mode application software'Division'is developed,by modifying and optimizing the arithmetic.It achieves good image division quality and will be helpful for doctor to make diagnosis.