Precise Feature Extraction for Multimodal Biometric Authentication Using Ranking Based CNN

Abstract

Multimodal biometric techniques have attracted increasing the recognition performance in some difficult biometric problems. A biometric system is a pattern recognition system that uses biometric characteristics to identify the people. Single biometric trait could not be sufficient to satisfy the biometric systems growing need for high precision. Multimodal biometric system can be used to increase the performance metrics that may not be possible using single biometric traits. Single biometric traits having difficulties such as disease infections, injuries, and challenges at a larger distance may harm the result. For the purpose of verifying an individual identification, there are currently available a variety of techniques. It is possible to identify someone based on their fingerprints, faces, signatures, voice and even the appearance of their irises when they gaze into the camera. newlineThe major goal of this research work is to create a multimodal biometric system that uses deep learning to recognize individuals based on their iris, palm print and lip biometric traits using MATLAB. There is currently limited research on the combination of these three traits. The third trait, Lip, has been considered to enhance the accuracy of the identification result and reliability of the proposed model. Deep Learning techniques extract certain characteristics from the input traits that can distinguish between wide range of people newline