Handwritten Digits Recognition With Subgroup Neural Network

作     者：WangMinghui PanXinan 等 

作者机构：DepartmentofInformationEngineeringBeijingUniversityofpostsandTelecommunicationBeijing100088P.R.China DepartmentofInformationEngineeringBeijingUniversityofpostsandTelecommunicationBeijing100088P. 

出 版 物：《The Journal of China Universities of Posts and Telecommunications》 (中国邮电高校学报（英文版）)

年 卷 期：1994年第1卷第1期

页      面：52-56页

核心收录：

学科分类：030604[法学-侦查学] 03[法学] 08[工学] 0838[工学-公安技术] 0306[法学-公安学] 030609[法学-涉外警务学] 

主　　题：Neural network Subgroup network Pattern recognition. 

摘      要：Most of the neural network handwritten digits recognition systems adopt a network with single structure. In this paper,we implement a subgroup neural network structure based on feature description to recognize handwritten digits. It is noticeable that this subgroup network is a hybrid system. The experiment result indicates that the subgroup network gets much better performance than usual single network Structure.