基于BP神经网络的手写数字识别系统  

Handwritten Digit Recognition System Based on BP Neural Network

作　　者：黄一格 张炎生 HUANG Yige;ZHANG Yansheng(School of Electronics and Information Engineering,Guangdong Ocean University,Zhanjiang,Guangdong 524088,China)

机构地区：[1]广东海洋大学电子与信息工程学院

出　　处：《机电工程技术》2020年第1期108-110,共3页Mechanical & Electrical Engineering Technology

摘　　要：针对多数机构面临的大规模报表数据录入问题,提出了一种基于BP神经网络的手写数字识别系统。对输入图像进行图像预处理、图像分割和特征提取,随后将提取的特征信息输入到已经训练好的BP神经网络进行分类识别。训练数据包含140幅大小归一的数字图片,其中100幅作为训练集,40幅作为验证集,并以10幅带有若干手写体数字的图片作为测试集进行识别分析。经Matlab仿真实验结果表明,该分类器具有较短的收敛时间和较为理想的识别精度,在实际工作中具有良好的应用价值。Aiming at the large-scale report data entry problem faced by most organizations,a handwritten digit recognition system based on BP neural network was proposed. image preprocessing,image segmentation and feature extraction were performed on the input image,and the extracted feature information was input into the trained BP neural network for classification and recognition. The training data consisted of 140 spoke-size digital pictures, with 100 spokes as the training set, 40 spokes as the validation set, and 10 spokes with several handwritten digits as the test set for recognition analysis. The simulation results of Matlab show that the classifier has short convergence time and ideal recognition accuracy,and has good application value in practical work.

关 键 词：手写数字识别 BP神经网络 图像预处理 特征提取 分类模型 识别精度 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]