Communication Establishment Based on Authenticating Earprints

Abstract

Security systems widespread and play essential roles in our lives. The strongest security systems can be built by biometric traits. Earprint can be considered as one of the most important biometrics, especially for phone communications. In this paper, preprocessing steps are adopted for segmenting the used image and extracting its earprint pattern. In addition, a Deep Learning (DL) model has been suggested for personal verification. It is called the Normalized Deep Earprint Learning (NDEL). Furthermore, a communication prototype has been designed to analyze an earprint, then, provide a secure mobile phone call. The Earprint Image of Northern Technical University (EINTU) database has been utilized. Efficient preprocessing steps are successfully suggested for segmenting earprints. The results show that the NDEL achieved 98.0% and 99.7% for recognizing persons by using their employed left and right earprint patterns, respectively. Finally, secure mobile phone calls are achieved by utilizing earprint authentications.