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

作　　者：郑蕊蕊[1] 赵继印[1] 李敏[1] 于为民[1] 吴宝春[1] 

机构地区：[1]大连民族学院信息与通信工程学院,辽宁大连116605

出　　处：《大连民族学院学报》2011年第5期502-506,共5页Journal of Dalian Nationalities University

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)

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。An offline Chinese handwritten characters and text database, DNU - I multi - 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 - I database. The DNU - I database consists of 3 subsets, the single character dataset, the single line dataset and the paragraph dataset. Each sample of the DNU - I database 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 - I database 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.

关 键 词：脱机手写体汉字识别 数据库 少数民族 图像处理 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]