Ear-biometrics for human identification

Abstract

The potential of human ear for identification was advocated long ago. It has been proved that ear of every individual is unique and can be used as a biometric to overcome the limitations of the biometrics used today. This paper presents an approach towards using ear as a biometric for identification. In this paper, a deep learning based, Convolutional neural network model is applied to the digitally processed database which gives promising results. We have used Gaussian filter and Canny edge detector for processing the image to increase recognition rate. For authentication we have used database provided by University Of Science and Technology(USTB), Bejing.