Performance Improvement of Multibiometric Authentication System Using Score Level Fusion with Ant Colony Optimization

Abstract

Biometric systems are widely used by many organisations to protect the data from anomaly users. Unimodal biometric systems have many problems like noisy data, nonversatility, and nonuniversality. To avoid these problems, multibiometric system is the most suitable approach where we can integrate more than two individual modalities. Our proposed framework is utilised to minimize the rate of error while working on the exhibition by utilising the methodology of Ant Colony Optimization in view of Score Level Fusion strategies. The proposed work will extract the highlights from two distinct modalities of individual people like iris and face. The proposed frameworks employ ACO as an optimization technique to choose the fusion parameters called weight to apply the fusion rule for different biometric matcher used for Score Level Fusion mechanism. The matching scores will be calculated based on the fusion methods like sum, tanh, mean, median, min, max, and product. Our proposed system implements and analyzes four different types of fusion mechanisms.