吉林大学学报(理学版)

• 计算机科学 • 上一篇    下一篇

改进特征匹配算法在银行卡号识别中的应用

董延华, 陈中华, 宋和烨, 黄雨   

  1. 吉林师范大学 计算机学院, 数值模拟吉林省高校重点实验室, 吉林 四平 136000
  • 收稿日期:2017-03-20 出版日期:2018-01-26 发布日期:2018-01-24
  • 通讯作者: 黄雨 E-mail:362395362@qq.com

Application of Improved Feature Matching Algorithmin Bank Card Number Recognition

DONG Yanhua, CHEN Zhonghua, SONG Heye, HUANG Yu   

  1. Jilin Province Key Laboratory for Numerical Simulation, College of Computer, Jilin Normal University, Siping 136000, Jilin Province, China
  • Received:2017-03-20 Online:2018-01-26 Published:2018-01-24
  • Contact: HUANG Yu E-mail:362395362@qq.com

摘要: 针对目前银行卡号识别算法的不足, 基于模式识别技术, 提出一种改进的特征匹配算法, 并将其应用于银行卡号识别. 改进的特征提取算法改变了特征向量, 并增加特征学习次数, 解决了传统方法识别银行卡号正确率较低的问题, 并保证了识别效率. 仿真实验结果表明, 改进的特征提取算法可有效提高银行卡号识别的正确率.

关键词: 改进特征匹配, 欧氏距离, 特征向量, 数字识别

Abstract: Aiming at the shortcomings of current bank card number recognition algorithm, we proposed an improved feature matching algorithm based on pattern recognition technique, and applied it to bank card number recognition. The improved feature extraction algorithm transformed feature vectors and increased the number of feature learnings,  solved the problem of traditional method identifying bank card number with low accuracy, and ensured the recognition efficiency. Simulation experimental results show that the improved feature extraction algorithm can effectively improve the correct rate of bank card number recognition.

Key words: Euclidean distance, improved feature matching, feature vector, numerical recognition

中图分类号: 

  • TP391.4