基于RBF神经网络的医学图像分类算法研究

摘要

This paper presents a medical image classification algorithm based on RBF neural network using feature information of pixels as training samples to train the RBF neural network , trained in the use of RBF neural network image classification and recognition , classification of different results give different RGB values are displayed.Experimental results show that by using RBF neural network , users not only can effectively distinguish between the medical images of different structure and displaying the image details, and the error curve can stable and fast convergence.%本文提出一种基于RBF神经网络的医学图像分类算法,利用像素的特征信息作为训练样本对RBF神经网络进行训练,使用训练后的RBF神经网络图像的分类识别,对不同的分类结果赋予不同rgb值进行显示.实验结果表明,RBF神经网络的结构简单、学习权值数量少且收敛速度快,不仅能有效的区分医学图像中的不同结构,显示图像细节,同时其误差曲线收敛稳定、速度快.