计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (31): 172-174.

• 数据库与信息处理 • 上一篇    下一篇

一种图片分类的序列构造神经网络方法

王仁武1,2,陈家训1   

  1. 1.东华大学 信息学院,上海 200051
    2.华东师范大学 信息学系,上海 200062
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-11-01 发布日期:2007-11-01
  • 通讯作者: 王仁武

Method of classification for images based on SCNN

WANG Ren-wu1,2,CHEN Jia-xun1   

  1. 1.College of Information,Donghua University,Shanghai 200051,China
    2.Department of Information,East China Normal University,Shanghai 200062,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-11-01 Published:2007-11-01
  • Contact: WANG Ren-wu

摘要: 针对图片数据的分类问题,从动态构造网络的思想出发,讨论了序列构造神经网络的基本方法,并且把其应用于图片分类中,最后给出标准测试集的分类测试结果,并对其进行了比对、分析和讨论,实验结果表明此方法适合多维数据分析,取得了满意的效果。

Abstract: To the problems of image classification,Sequence Constructive Neural Network(SCNN) method is discussed using the
idea of dynamic creation of neural network and its application in the image classification.Benchmark tests are launched in the paper,and also give its compares and analysis.From the experiments,the method of SCNN is more suitable for multi-dimension data analysis and can give satisfied results.