首页 | 本学科首页   官方微博 | 高级检索  
     检索      

浅谈神经网络的图像识别技术及方法
引用本文:陈华.浅谈神经网络的图像识别技术及方法[J].济宁师范专科学校学报,2006,27(6):20-22.
作者姓名:陈华
作者单位:济宁师范专科学校计算机系,山东曲阜273155
摘    要:神经网络图像识别技术是一种新型图像识别技术,是在传统的图像识别方法和基础上融合神经网络算法的一种图像识别方法。神经网络用于图像识别一般都要提取图像特征,然后把提取好的图像特征送入神经网络识别器进行识别。BP神经网络图像识别方法不对图像作特征提取,直接把图像数据作为神经网络识别器的输入。通过用MATLAB完成的网络的训练与测试表明,不作特征提取的神经网络图像识别系统具有很强的抗干扰能力。

关 键 词:图像识别  人工神经网络(ANN)  BP网络算法
文章编号:1004-1877(2006)06-0020-03
收稿时间:2006-11-03
修稿时间:2006年11月3日

Introduction of the image Recognition technology and methodology in neural network
CHEN Hua.Introduction of the image Recognition technology and methodology in neural network[J].Journal of Jining Teachers College,2006,27(6):20-22.
Authors:CHEN Hua
Institution:Department of Computer of Jining Normal College, Qufu 273155 ,China
Abstract:the image recognition technology and methodology of neural network is a kind of image recognition technology. It is a kind of technology which based on the traditional image recognition technology and work with the neural network arithmetic together. Neural network is used for image recognition. General, we pick - up the image characteristics firstly, and then input the image characteristics into the recognizer to recognize. However, BP neural network image recognition wont pick -up the image characteristics. Instead, it input the image to the recognizer directly. The testing for the MATLAB network indicates that the neural network image recognition system without a pick - up of characteristic will be more anti - jamming.
Keywords:MATLAB
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号