您的位置:首页  > 论文页面

基于屋脊滤波的指横纹认证

发表时间:2010-11-30  浏览量:1502  下载量:659
全部作者: 张延强,孙冬梅,裘正定
作者单位: 北京交通大学信息科学研究所
摘 要: 提出基于屋脊边缘滤波的指横纹定位与提取算法,采用基于归一化互相关方法对特征点进行匹配。在98人、1 971幅图像的测试数据上表明指横纹作为生物特征的普遍性和有效性。单模最优系统为无名指指横纹,等错误率(equal error rate, EER)为0.877 8%;最差系统为小拇指指横纹,EER为2.040 3%. 采用fisher判别法融合4类指横纹,系统EER为0.207 8%,达到了较高的安全级别。
关 键 词: 信号与信息处理;指横纹;屋脊滤波;互相关点匹配;决策级融合
Title: Inner knuckleprint verification based on roof edge detector
Author: ZHANG Yanqiang, SUN Dongmei, QIU Zhengding
Organization: Institute of Information Science, Beijing Jiaotong University
Abstract: A novel biometrics defined as ‘inner knuckleprint’ is presented in this paper. Roof edge detector is used for locating and extracting the inner knuckleprint. Normalized cross-correlation based feature matching and decision fusion scheme are integrated to implement a real-time verification system. The system is evaluated based on the database contains 1 971 image samples from 98 individuals. For single model, the equal error rate (EER) is 2.040 3% to the worst system, while 0.877 8% to the best one, which indicates that the inner knuckleprints are reliable and universal as a biometrics, and demonstrates the effectiveness of the proposed method.
Key words: signal and information processing; inner knuckleprint; roof edge detector; normalized cross-correlation matching; decision level fusion
发表期数: 2010年11月第22期
引用格式: 张延强,孙冬梅,裘正定. 基于屋脊滤波的指横纹认证[J]. 中国科技论文在线精品论文,2010,3(22):2266-2273.
 
0 评论数 0
暂无评论
友情链接