节点文献

多民族脱机手写体汉字数据库的设计与构建

DNU-Ⅰ Multi-national Offline Chinese Handwritten Database of Dalian Nationalities University

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 郑蕊蕊赵继印李敏于为民吴宝春

【Author】 ZHENG Rui-rui,ZHAO Ji-yin,LI Min,YU Wei-min,WU Bao-chun(College of Information & Communication Engineering,Dalian Nationalities University,Dalian Liaoning,116605,China)

【机构】 大连民族学院信息与通信工程学院

【摘要】 设计并构建了一种记录书写者民族信息的手写体汉字数据库———大连民族学院DNU-Ⅰ型多民族脱机手写体汉字数据库。包括单字库、行文本库和段文本库3个子库。为少数民族汉字书写特征分析、中文文档的行切分、汉字的切分识别、中文文本的无切分识别、笔迹鉴别和签名验证等方面的研究奠定基础,并提供算法的验证平台。同时介绍了字符识别数据库的一般构建流程和数据库图像二值化、归一化、行分割等预处理算法,为少数民族文字数据库的构建提供了技术支撑。

【Abstract】 An offline Chinese handwritten characters and text database,DNU-Ⅰmulti-national offline Chinese handwritten database of Dalian Nationalities University,has been presented to record the writers’ national information.Dalian Nationalities University has the copyright of the DNU-Ⅰdatabase.The DNU-Ⅰdatabase consists of 3 subsets,the single character dataset,the single line dataset and the paragraph dataset.Each sample of the DNU-Ⅰdatabase recorded the writer’s information,such as his or her name,nationality,gender and education.The proportion of writers from minority nationalities is 60%.The DNU-Ⅰdatabase can be used to conduct written features of minority nationalities,Chinese text line segmentation,Chinese characters segmentation,segmentation-free recognition,writer identification,signature verification and provide benchmark for algorithms comparison.Meanwhile,common construction procedures of character recognition database and the binarization,normalization,and line segmentation methods of character image pre-processing,which can provide technique support for minority nationalities’ written languages,has been introuduceed.

【基金】 国家科技支撑计划项目(2009BAH41B05);国家民委科研项目(10DL03);辽宁省教育厅项目(L2010094);中央高校基本科研业务费专项资金资助项目(DC10010103);大连民族学院人才引进科研启动基金资助项目(20116203)
  • 【文献出处】 大连民族学院学报 ,Journal of Dalian Nationalities University , 编辑部邮箱 ,2011年05期
  • 【分类号】TP391.43
  • 【被引频次】1
  • 【下载频次】94
节点文献中: 

本文链接的文献网络图示:

本文的引文网络