基于RBF神经网络的医学图像分类算法研究  

The classification of medical image based on the RBF neural network

作　　者：周慧[1] 张尤赛[1] 龚淼[1] ZHOU Hui ZHANG You-sai GONG Miao(Department of Electronics and Information, Jiangsu University of Science and Technology, Zhenjiang 212003, China)

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212000

出　　处：《电子设计工程》2017年第3期113-116,120,共5页Electronic Design Engineering

摘　　要：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。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.

关 键 词：医学图像 特征信息 RBF神经网络 图像分类 

分 类 号：TN911.73[电子电信—通信与信息系统]