Personal Verification Using Fast and Self Consistent Estimation for Biometric Match Score Fusion

摘要

In this paper, the authors propose a novel personal verification system that combines face and fingerprint features. In the proposed system, face and fingerprint features are extracted by Zernike Moment (ZM); the score densities after the matching are fused based on fast and self-consistent estimation (FSCE) method that is used for estmating the genuine and impostor densities for personal verification. Experimental results on the FVC2004 and Yale databases show that FSCE provides high accuracy as well as achieves low error rates. Moreover, the comparison results suggest that the proposed score level fusion of multimodal using FSCE outperforms other the state-of-the-art approaches (such as GMM, SVM).