一种基于虹膜和人脸的多生物特征融合方法

针对单一生物特征识别方法易受干扰、应用受限制等问题,提出了一种基于虹膜和人脸的多生物特征融合方法以提高身份识别的精度及可靠性.该融合方法利用 Log-Gabor 相位编码算法和Laplacianfaces 算法对虹膜和人脸进行特征提取及匹配,然后在匹配层对两种生物特征的匹配得分值使用最小最大概率机策略进行融合,最后利用融合后得分值进行决策.该方法与单生物特征识别方法相比,识别性能明显提高.在UBIRIS虹膜库和ORL人脸库构成的多模生物特征库中进行了测试,实验结果表明:该方法的等错误率可降低到 0.28%,比单一虹膜和单一人脸方法分别降低了0.69%和 1.62%,与采用传统融合策略的方法相比,等错误率也降低了 0.2%以上.

Multimodal Biometric Fusion Approach Based on Iris and Face

The recognition performance of the single biometric system is easily affected by the surrounding environment.In order to improve the recognition reliability,a multimodal biometric fusion approach based on iris and face was proposed.In the proposed approach,the iris is processed using the phase information algorithm based on Log-Gabor filtering and the face is processed using Laplacianfaces algorithm.The two matching scores obtained from iris and face are fused using the minimax probability machine fusion strategy,and the fused matching score is used to make the final decision.The performance of the proposed approach is obviously improved.The experimental results in a multimodal database,which consists of UBIRIS iris database and ORL face database,show that the equal error rate of the proposed approach is 0.28%,which decreases 0.69% and 1.62%,respectively compared with both the single biometric approaches and decreases over 0.2% compared to the tradition fusion strategies.