首页 | 本学科首页   官方微博 | 高级检索  
     

采用多级分类器的手写数字识别技术研究
引用本文:赵锋,薛惠锋,王伟,吴慧欣. 采用多级分类器的手写数字识别技术研究[J]. 微电子学与计算机, 2006, 23(12): 30-32,37
作者姓名:赵锋  薛惠锋  王伟  吴慧欣
作者单位:1. 西北工业大学,自动化学院,陕西,西安,710072
2. 西北工业大学,航空学院,陕西,西安,710072
摘    要:在分析目前手写数字识别策略基础上,提出了采用多级分类器的手写数字识别技术,并给出了多级分类器识别系统体系结构。首先对手写数字样本进行预处理,然后针对不同的子分类器,提取不同的分类特征。系统中上一级的拒识样本作为下一级的输入样本,因此系统识别率逐级增加。最后以面向对象语言Visual C++和MatLab为工具.开发了多级分类器实验原形系统,结果表明,该识别技术不仅操作简单,而且具有较好的识别性能。

关 键 词:分类器  特征抽取  二值化  数字识别
文章编号:1000-7180(2006)12-0030-03
收稿时间:2005-11-21
修稿时间:2005-11-21

Handwritten Digit Recognition Technology Based on Multistage Classifying Device
ZHAO Feng,XUE Hui-feng,WANG Wei,WU Hui-Xin. Handwritten Digit Recognition Technology Based on Multistage Classifying Device[J]. Microelectronics & Computer, 2006, 23(12): 30-32,37
Authors:ZHAO Feng  XUE Hui-feng  WANG Wei  WU Hui-Xin
Abstract:On the basis of analyzing present Handwritten Numeral Recognition,a multistage classifying device is proposed and a multistage Classifying device system structure is given.Tow procedure are included:Preconditioning on Handwritten Digital samples;Drawing different classify characters in light of different subclassfying device.The identifying ratio increase by level number as the refuse identify samples of the last level are the input of the next level.Finally,a multistage classifying device experimental system is developed based on Visual C++ and Matlab.The result demonstrates that the identifying technology.Operates simply and has better performance as well.
Keywords:Classifying device   Feature choice   Two values   Handwritten numeral recognition
本文献已被 CNKI 维普 万方数据 等数据库收录!
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