| 多分类器融合在银行票据识别中的应用 |
| |
| 引用本文： | 朱辉,唐降龙,孙广玲. 多分类器融合在银行票据识别中的应用[J]. 计算机工程与应用, 2003, 39(30): 219-222 |
| |
| 作者姓名： | 朱辉 唐降龙 孙广玲 |
| |
| 作者单位： | 哈尔滨工业大学计算机科学与技术学院,哈尔滨,150001 |
| |
| 摘 要： | 论文将数字识别技术应用于银行票据自动识别勾兑业务系统,并利用多分类器融合方法获得了99.527%的单个数字字符识别率,在此基础上运用多区域多内容冗余校验方法保证了识别结果的充分可信赖性,经此处理后整张票据的识别通过率高达87.573%,其识别结果完全正确。
|
| 关 键 词： | 多分类器组合 银行票据识别 |
| 文章编号： | 1002-8331-(2003)30-0219-04 |
| 修稿时间： | 2002-08-01 |
Application of Multiple Classifiers Fusion in Bank Paper Recognition |
| |
| Zhu Hui Tang Jianglong Sun Guangling. Application of Multiple Classifiers Fusion in Bank Paper Recognition[J]. Computer Engineering and Applications, 2003, 39(30): 219-222 |
| |
| Authors: | Zhu Hui Tang Jianglong Sun Guangling |
| |
| Abstract: | In this paper methods of multiple classifiers fusion are used in bank paper recognition system to increase single character's recognition rate.The result shows that single character recognition rate can reach to99.527%.Based on this high recognition rate,the authors recognize multiple regions and use their results as redundant information to verify recognized results.87.573%pass rate of papers whose recognized results are totally correct can be obtained. |
| |
| Keywords: | Multiple classifiers fusion bank paper recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
