摘要
针对传统手写汉字识别特征提取过程复杂,识别率低及通用深度学习分类模型判别能力较弱的问题.本文设计了一种多通道交叉融合的深度残差网络模型并对中心损失函数做出了改进.首先,通过对原始数据集进行预处理来降低模型过拟合的风险;然后,将经本文改进后的中心损失函数和Softmax损失函数联合作为模型训练的监督信号,在训练过程中有效的使数据集类内聚合、类间分散,提高了模型的分类性能;最后,将经过预处理的数据集输入到本文设计的模型中,通过多次训练进行参数调整得到最佳识别效果.在CASIAHWDB-V1. 1数据集上的实验表明本文设计的识别模型及算法有效的提高手写汉字的识别率.
Aiming at the complex feature extraction process of traditional handwritten Chinese character recognition,lowrecognition rate and weak discrimination ability of general deep learning classification model,a multi-channel cross-fusion deep residual network model and improves the central loss function designed in this paper. Firstly,based on the original data set pre-processing to reduce the risk of the model in over-fitting;then,the improved central loss function and the Softmax loss function are combined as the monitoring signal of model training,which can effectively aggregate data sets within the class,disperse between classes and improve the classification performance of the model in the training process;finally,the pre-processed data sets are input into the model designed in this paper. The experiment on CASIAHWDB-V1. 1 data set show s that the recognition model and algorithm designed in this paper can effectively improve the recognition rate of handwritten Chinese characters.
作者
张秀玲
周凯旋
魏其珺
董逍鹏
ZHANG Xiu-ling;ZHOU Kai-xuan;WEI Qi-jun;DONG Xiao-peng(Key Laboratory of Industrial Computer Control Engineering of Hebei Province,Yanshan University,Qinhuangdaoi 066004,China;National Engineering Research Center for Equipment and Technology of Cold Strip Rolling,Yanshan University,Qinhuangdao 066004,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2019年第10期2232-2235,共4页
Journal of Chinese Computer Systems
基金
河北省自然科学基金项目(E2015203354)资助
河北省高校创新团队领军人才培育计划项目(LJRC013)资助
河北省教育厅科学研究计划河北省高等学校自然科学研究重点项目(ZD2016100)资助
2016年燕山大学基础研究专项培育课题项目(16LGY015)资助
河北省高等教育学会高等教育科学研究课题重点课题项目(GJXHZ2015-1)资助
关键词
交叉融合
深度学习
残差网络
中心损失
手写汉字识别
cross fusion
deep learning
resnet
center loss
handwritten Chinese character recognition