摘要
联机手写汉字识别的模拟仿真系统是基于卷积神经网络方法开发的。利用TensorFlow实验平台,以中国科学院采集的CASIA-OLHWDB数据集进行训练获得分类数据。使用基于Python的Flask框架搭建Web服务器,采用Ajax技术实现数据在Web端的实时自动刷新。仿真结果表明,本系统能够较精确识别联机手写汉字,达到预计目的。
An online handwritten Chinese character recognition simulation system was developed on the basis of the algorithm of convolution neural network,in which classified data were obtained by drilling with the CASIA-OLHWDB dataset collected by Chinese Academy of Science on TensorFlow platform and refreshed real-time and automatically via Ajax on the Web servers built upon the Python-based Flask framework.The result shows that the system is able to recognized online handwritten Chinese characters accurately enough to meet the requirements.
作者
曲丽娜
QU Li-na(School of Information Engineering at Jilin Engineering Normal University,Changchun Jilin 130052,China)
出处
《吉林工程技术师范学院学报》
2018年第10期105-107,共3页
Journal of Jilin Engineering Normal University
基金
吉林省教育厅"十二五"科学技术研究项目(2014571)
