People Recognition Based on Fusion at The Score Level Using NISTDatabase

Abstract

The identification of individuals has become essential to ensure the security of systems and organizations. Biometrics refers to the automatic recognition of individuals based on their physiological or behavioral characteristics. However, the unimodal biometric system suffers from certain limitations, to overcome these problems and to get better performance in terms of recognition rate, information from different biometric sources is combined, and these systems are called multimodal biometric systems. Integration into multimodal biometric systems takes place at four levels and score level fusion is the most effective. In this paper, a scheme for score-level fusion has been proposed based on different methods of fusion the score (usual and proposed). Experimental results on one of the three databases accessible to the public of NIST-BSSR 1, this database data called NIST-BSSR1 Set2 (NIST-Fingerprint), The results obtained showed the superiority of the proposed approach over many of the fusion methods proposed by the researchers.