Feature level fusion of iris and finger vein biometrics for multimodal biometric authentication system

Abstract

With the intense need of security, a reliable authentication system can be attained using multimodal biometrics. Predominately vein patterns are attracting the researchers for developing authentication system. Multimodal biometric system not only aims at combining traits but also on fusion at various levels. Proposed approach fuses invariant iris features and finger vein shape features. The fusion at feature level framework is evaluated to perceive classification accuracy of biometric authentication system. Algorithm prioritizes on reducing high dimension features by considering iris Hu moments and finger vein shape features to accomplish a secured and convenient authentication system. SVM Classifier results prove that multimodal biometric outperforms compared to Uni-modal system.