首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于多尺度梯度及深度神经网络的汉字识别
引用本文:潘炜深,金连文,冯子勇.基于多尺度梯度及深度神经网络的汉字识别[J].北京航空航天大学学报,2015,41(4):751-756.
作者姓名:潘炜深  金连文  冯子勇
作者单位:华南理工大学电子与信息学院,广州,510641;华南理工大学电子与信息学院,广州,510641;华南理工大学电子与信息学院,广州,510641
基金项目:国家自然科学基金资助项目,国家科技支撑计划资助项目,广东省科技计划资助项目
摘    要:介绍了一种基于多尺度滑动窗的方法提取文字的梯度直方图特征,并结合深度神经网络对印刷体汉字进行识别.针对梯度直方图的空间关系,使用可伸缩的滑动窗对图像进行分割,在不同尺度上获取文字的特征信息,有效融合汉字的全局特征和局部分块特征.实验采用5层的深度神经网络模型对国标一级3 755个印刷体汉字进行分类,并应用Dropout技术防止训练过拟合,提高神经网络的泛化能力.实验准确率达到98.292%,有较好的识别性能,验证了本文多尺度梯度特征及深度神经网络模型在文字识别上的有效性.

关 键 词:多尺度滑动窗  梯度直方图  深度神经网络  泛化能力  汉字识别
收稿时间:2014-04-28

Recognition of Chinese characters based on multi-scale gradient and deep neural network
PAN Weishen , JIN Lianwen , FENG Ziyong.Recognition of Chinese characters based on multi-scale gradient and deep neural network[J].Journal of Beijing University of Aeronautics and Astronautics,2015,41(4):751-756.
Authors:PAN Weishen  JIN Lianwen  FENG Ziyong
Abstract:The method to extract the gradient histogram feature of the Chinese characters with a multi-scale sliding window and to recognize the printed Chinese characters with deep neural network was presented. In order to acquire the spatial information of the gradient histogram, a retractable sliding window technique was proposed for segmenting the images and getting the gradient feature information from different scales which can effectively combine all the global features and local block features of Chinese characters. The experiment was carried out by using a 5-layer deep neural network to classify 3755 categories of printed Chinese characters.A Dropout technique was applied so as to prevent over-fitting training and to improve the generalization ability of the neural network. The accuracy of the experiment reaches 98.292%, which has better recognition performance and demonstrates that the method of applying a multi-scale gradient feature and deep neural network model on the recognition of Chinese characters is effective.
Keywords:multi-scale sliding window  gradient histogram  deep neural network  generalization ability  recognition of Chinese characters
本文献已被 CNKI 万方数据 等数据库收录!
点击此处可从《北京航空航天大学学报》浏览原始摘要信息
点击此处可从《北京航空航天大学学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号