Convolutional Neural Network-based Finger Vein Recognition using Near Infrared Images

Abstract

Convolutional Neural Network (CNN) is opening \nnew horizons in biometrics-based authentication field and \nfinger vein recognition is the prominent one which can provide \nthe best possible security system depending on this \naforementioned technology. In this paper, we used 5 \nconvolutional layers and 4 fully-connected layers where our \ndeveloped network has shown the capability to produce the \nresult with almost 100% accuracy rate which became possible \ndue to the fact that deep learning, an end-to-end system is used \nwhich performs better in a lot of aspects in comparison to \nconventional techniques.Convolutional Neural Network (CNN) is opening \nnew horizons in biometrics-based authentication field and \nfinger vein recognition is the prominent one which can provide \nthe best possible security system depending on this \naforementioned technology. In this paper, we used 5 \nconvolutional layers and 4 fully-connected layers where our \ndeveloped network has shown the capability to produce the \nresult with almost 100% accuracy rate which became possible \ndue to the fact that deep learning, an end-to-end system is used \nwhich performs better in a lot of aspects in comparison to \nconventional techniques.