基于人工神经网络的手写体数字识别  

Recognition for Handwriting Numbers Based on Artificial Neural Network

作　　者：邓丽华[1] 崔志强[1] 张静[2] 

机构地区：[1]三峡大学电气信息学院,湖北宜昌443002 [2]湖北三峡职业技术学院机电工程系,湖北宜昌443000

出　　处：《三峡大学学报（自然科学版）》2005年第3期254-256,共3页Journal of China Three Gorges University：Natural Sciences

摘　　要：基于手写体数字的几何结构提取端点及其方向、凸点及其方向、三叉点数和四叉点数等数字特征值,运用单字单网的12个并行BP神经网络进行数字识别,把同一个数字特征值差别很大的字体当不同的类别来处理,最后能达到很好的识别效果.Based on free handwriting numbers geometric structure, a way about picking up numbers eigenvalue for end and its position, protruding point and its direction, three cross point's position, four cross point's position and so on, is put forward; and number which exerts twelve paratactic BP neuron network for single number and single network, is recognized; and regard the different scripts of a same number, which differ significantly from characteristic value, as different classification, therefore it obtains better effect.

关 键 词：手写体数字识别 人工神经网络 几何结构 系统结构 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]