A Feasible Adaptive Fuzzy Genetic Technique for Face, Fingerprint, and Palmprint Based Multimodal Biometrics Systems

Abstract

A biometric system relies solely on one or a few biometric characteristics to verify a person's identity. Multimodal biometric authentication is a hot emerging area of research. The memory requirements, response times, and adoption/operating costs of conventional multimodal biometric identification methods are all higher than those of single-modal approaches. In this article, we conducted an examination of a framework for multimodal biometric identification systems, which demonstrates a practical implementation of soft computing strategies adaptable to face, finger, and palmprint biometrics. We applied a modified Gabor filter for feature extraction to increase processing speed and reduces the timing. Validation of the proposed system was achieved by the development of a fusion system using principal component analysis as a single matcher classifier. An adaptive fuzzy genetic algorithm was applied for weight optimization which generates verification at a high-rate performance using the fuzzy logic function. Employing fusion in identification mode, the technology was critically examined. The results indicated that the multimodal biometric system outperforms in terms of TPR, FPR, TNR, FNR, Precision, Recall, F-score, and Accuracy, resulting in reduced processing time and memory footprint, and speedier implementation.