基于决策级融合的掌纹掌静脉多特征融合匹配算法研究
首发时间:2012-05-04
摘要:针对单一模态的生物特征识别系统存在固有的缺点,本文提出一种新的基于决策级融合的掌纹掌静脉多特征融合识别的方法。首先,对掌纹及掌静脉两种模态分别进行初级分类器设计,用多种匹配算法决策融合,得出识别结果;然后将提取出的掌纹与掌静脉特征点集,以串联的方式形成新的特征,构成二级分类器,再用同样的方法,得出一个识别结果;最后将上述两个识别结果再进行决策级融合,得到系统最后的结果.实验结果证明,本算法能有效的提高系统的通过率,降低误识率。
For information in English, please click here
Decision Fusion Of Palm And Plam Vein Matching Algorithm Based on Multiple Feature Fusion
Abstract:For a single modal biometric system have shortcomings, this paper proposes a new palm vein-based decision level fusion of palm multi-feature fusion recognition method. First, the two modes of palm and palm vein were the primary classifier design, a variety of matching algorithms decision fusion, obtained recognition results; then extracted palm and palm vein feature points set in series the formation of new features, constitute the two classifiers, and then in the same way, and come to a recognition result; the last of the two recognition results for decision level fusion, and to get the final result of the system. Experimental results show that this algorithm can effectively improve the system by reducing the error rate.
Keywords: Palm And Palm Vein Fusion Multi-feature Fusion Decision Level Fusion
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于决策级融合的掌纹掌静脉多特征融合匹配算法研究
评论
全部评论0/1000