基于PSO与BP神经网络的脱机手写体汉字识别算法  

Offline Handwritten Chinese Character Recognition Algorithm Based on PSO and BP Neural Network

作　　者：岳中彤 Yue Zhongtong(School of Electronic Communication and Physics, Shandong University of Science and Technology, Qingdao 266590, Chin)

机构地区：[1]山东科技大学电子通信与物理学院,青岛266590

出　　处：《信息化研究》2018年第2期68-70,共3页INFORMATIZATION RESEARCH

摘　　要：汉字识别的算法研究是模式识别中的热点课题。文章针对脱机手写体汉字提出了一种用PSO算法优化BP神经网络的脱机识别算法。关于BP算法在训练时易出现局部极小化的现象,PSO算法可通过大空间内的搜索能力,在全局中优化BP算法。文章基于粒子群算法优化BP神经网络(PSO-BPNN)研究脱机手写体汉字识别算法,通过Matlab软件对样本数据进行分类仿真。结果表明,PSO优化后的算法具有较高的收敛速度和稳定性,对手写体汉字的识别具有较强的能力。The algorithm research of Chinese character recognition is a hot topic in pattern recognition.This paper presents an offline recognition algorithm for off-line handwritten Chinese characters using PSO algorithm to optimize BP neural network[1].As the BP algorithm is prone to local miniaturization during training,the PSO algorithm can optimize the BP algorithm in the global through the search capability in large space.In this paper,off-line handwritten Chinese character recognition algorithm is studied based on PSO-BPNN,and the sample data is classified and simulated by Matlab software.The results show that the PSO optimized algorithm has higher convergence speed and stability.The algorithm has strong ability to recognize handwritten Chinese characters.

关 键 词：PSO算法 BP神经网络 粒子群算法优化BP神经网络 手写体汉字识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]