Multi-algorithmic Palmprint Authentication System Based on Score Level Fusion

Fusion of multiple algorithms utilizes as much information as possible from each algorithm for enhancing the performance of the biometric authentication system. It is a big challenge to formulate a single algorithm for any biometric authentication system to addresses the problem of illumination, orientations and pose variations. The palmprint features are extracted using two feature extraction algorithms namely contourlet transform with principal component analysis and dual-tree complex wavelet transform. The match scores obtained from matching modules were normalized using z-score and fused with different score level fusion schemes namely sum-rule,weighted sum-rule, and Support Vector Machine (SVM) fusion, respectively. The experimental results show that SVM score level fusion lead to an increased performance for the proposed multi-algorithmic palmprint authentication system with genuine acceptance rate of 98% for 0.1% false acceptance rate, and equal error rate of 1.5% compared to weighted sum-rule and sum-rule fusion schemes.