Design and development of 2D and 3D ear based recognition system using deep learning approaches

Abstract

Person identification using Human Ear image has been paid more attention now a days. Human ear are easily visible and their sample images are easily capture. The structure of human ear does not change drastically over the time and it satisfies the biometric traits like universality, uniqueness, stability and collectability. So it is considered as a good source of passive data for person identifications. For security applications, it plays great potential which is naturally a complement with the face recognition to strength the identification recognition. Recently, in the pandemic situation of COVID-19, identifying a person from their face becomes difficult due to wearing of mask and maintaining social distance. In this regard, ear play a significant role. There exist several biometric characteristics which fulfil the above requirements (like universality, distinctiveness, permanence and collectability). However there is no single biometric that can satisfy the needs of all applications. Many researchers focused on iris, gait, face, figure print, palm print, gesture, or hand- geometry. These biometric traits gives high level of uniqueness however the subjects cooperation is very importatnt in each cases for acceptability of recognition model. Recent works have shown that in the field of object detection, deep learning achieved very good result newline