基于kNN-SVM的手背静脉虹膜和指纹融合身份识别

Identification Algorithm for Fusion of Hand Vein Iris and Fingerprint Based on kNN-SVM

针对识别模式下多生物特征融合识别系统的实现问题,本文基于手背静脉、虹膜和指纹三种生物特征研究了高效的融合识别算法。分别对三种生物特征进行特征提取与匹配,得到独立的匹配分数,基于k近邻(k Nearest Neighbor,kNN)分类器实现手背静脉特征识别,将用户身份范围缩小到k个,实现个人身份的初步识别,利用支持向量机(Support Vector Machine,SVM)算法实现k个样本范围内虹膜和指纹的融合识别,实现最终的个人身份识别。利用构建的三模态生物特征图像数据库进行了实验分析,实验结果表明该系统具有较高的识别性能。

Aiming at the implementation of multi-modal biometric system in identification modal, the efficient identification algorithm based on hand vein, iris and fingerprint was developed. Firstly, feature extraction and feature matching of three unimodal biometric traits was carded out respectively and the independent matching scores of each trait was obtained. Then, k Nearest Neighbor （kNN） classifier was utilized to preliminary identification based on hand vein, and the number of user＇s identity would be reduced to k. Finally, Support Vector Machine （SVM） classifier was developed to accuracy identification of user＇s identity based on iris and fingerprint. The constructing three-modal biometric image database was used to experimental analysis. The results show that the system has good identification performance, which possesses wide application prospect.