Adaptive of the bee algorithm with an overall pattern extractor to improve the fingerprint matching process

Abstract

Fingerprints serve as a crucial biometric identifier for verifying individual identities, particularly in the context of developing automatic fingerprint recognition systems. This study addresses the imperative need to enhance fingerprint matching accuracy while concurrently mitigating false acceptance rate (FAR) and rejection rate (RR) challenges inherent in traditional matching methods. The research proposes a novel approach, integrating the adaptive bee algorithm with a global pattern extractor to refine fingerprint matching processes. Specifically, the methodology focuses on adjusting the central point and angle of the studied impression. The effectiveness of the proposed methodology was evaluated using the standard FVC2004 database.
Results demonstrate the efficacy of the bee algorithm, a prominent artificial intelligence technique within the realm of swarm intelligence, in achieving a remarkable matching rate of 97.3%. This outcome underscores the algorithm's potency compared to alternative methods employed for similar purposes.