基于多分类的虹膜匹配方法  

Method of Iris Matching Based on Multi-classification

作　　者：代刚[1] 张海滨[2] 

机构地区：[1]中国电子科技集团公司第五十四研究所,河北石家庄050081 [2]海军装备部驻天津地区军事代表局,天津300061

出　　处：《无线电工程》2013年第11期60-64,共5页Radio Engineering

摘　　要：在传统的虹膜识别系统中,虹膜匹配被认为是一个二分类问题:类内匹配和类间匹配。许多已存在的方法简单地利用距离来执行虹膜匹配。由于这些方法不能很好地利用虹膜特征,所以会产生很高的拒识率和误识率,且鲁棒性不强。为了解决这些问题,提出把虹膜匹配当作一个多分类问题,采用一种新颖的蕨算法(Ferns)分类器来完成该工作。相比支持向量机(SVM)分类器,在执行虹膜匹配时,Ferns分类器有诸多优点。为了对提出的算法给出全面评价,实验中分别在认证和识别这2种模式下对该算法进行测试。实验结果证明,提出的方法可以极大地改善虹膜识别系统的性能。In conventional iris recognition systems, the iris matching is considered as binary-classification problem such as authen- tic matching and imposter matching. Many existing methods employ simple distance to implement iris matching,such as Hamming dis- tance. These methods can＇t make full use of iris feature,which causes relatively high FRR （False Reject Rate） , FAR （False Accept Rate） and low robustness. To overcome these weak points,this paper treats iris matching as a multi-classification problem. Though SVM is usually used to implement multi-classification,this paper chooses Ferns classifier which has many advantages to replace SVM. To evaluate our proposed iris matching algorithm completely, it is tested in two modes such as verification and identification. The experimental results show that our method can improve performance of iris recognition system greatly.

关 键 词：虹膜识别 Ferns分类器 SVM 虹膜匹配 接受器工作特性(ROC)曲线 

分 类 号：TN391.4[电子电信—物理电子学]