MULTIMODAL BIOMETRIC IDENTIFICATION SYSTEM USING THE FUSION OF FINGERPRINT AND IRIS RECOGNITION WITH CNN APPROACH

Abstract

Multimodal biometric systems are widely applied in many real-world applications because of its ability to accommodate variety of great limitations of unimodal biometric systems, including sensitivity to noise, population coverage, intra-class variability, nonuniversality, and vulnerability to spoofing. during this paper, an efficient and real-time multimodal biometric system is proposed supported building deep learning representations for images of both the correct and left irises of someone, and fusing the results obtained employing a ranking-level fusion method. The trained deep learning system proposed is named IrisConvNet whose architecture relies on a mix of Convolutional Neural Network (CNN) and Softmax classifier to extract discriminative features from the input image with none domain knowledge where the input image represents the localized iris region and so classify it into one amongst N classes. during this work, a discriminative CNN training scheme supported a mixture of back-propagation algorithm and mini-batch AdaGrad optimization method is proposed for weights updating and learning rate adaptation, respectively. additionally, other training strategies (e.g., dropout method, data augmentation) also are proposed so as to gauge different CNN architectures. The performance of the proposed system is tested on three public datasets collected under different conditions: SDUMLA-HMT, CASIA-IrisV3 Interval and IITD iris database