基于BP神经网络的手写数字识别方法的实现  

Realization of handwritten numeral recognition method based on BP neural network

作　　者：张黎[1] 刘争鸣[1] 唐军[1] 

机构地区：[1]贵州大学电气工程学院,贵阳550025

出　　处：《自动化与仪器仪表》2015年第6期169-170,共2页Automation & Instrumentation

摘　　要：目前,人工神经网络由于其理论可靠、通用性好等优点而广泛应用在识别软件开发上,本文对手写输入数字识别进行了研究,提出一种基于BP神经网络的识别方法,通过将手写输入数字二值化进行辨识来增加神经网络的辨识成功率,实验结果表明,应用BP神经网络的手写数字识别方法可以有效识别各种写法的数字。Recently,Artificial Neural Network used wildly on Recognition system because of its reliable theory and Generality.In this paper,we put forward a recognize method based on BP neural network,and increased network recognition rate by binaryzation of handwritten numbers.The experiment result shows that the handwritten number recognition program based on BP neural network could recognize handwritten numbers of different writing effectively.

关 键 词：BP神经网络 手写字符识别 字符提取 

分 类 号：TP391.41[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]