Behavioral Biometric Authentication in Android Unlock Patterns through Machine Learning

José Torres, Sergio Santos, Efthimios Alepis, Constantinos Patsakis

Abstract

The penetration of ICT in our everyday lives has introduced numerous automations, and the continuous need for communication has made mobile devices indispensable. As a result, there is an ever-increasing deployment of services for which users need to authenticate. While the use of plain passwords is the default, many applications require higher standards of security, such as drawn patterns and fingerprints, used mostly to authenticate users and unlock their smart devices. In this work we propose a biometrics-based machine learning approach that supports user authentication in Android to augment native user authentication mechanisms, making the process more seamless and secure. Our evaluation results show very high rates of success, both for authenticating the legitimate user and also for rejecting the false ones. Finally, we showcase how the proposed solution can be deployed in non-rooted devices.

Download


Paper Citation


in Harvard Style

Torres J., Santos S., Alepis E. and Patsakis C. (2019). Behavioral Biometric Authentication in Android Unlock Patterns through Machine Learning.In Proceedings of the 5th International Conference on Information Systems Security and Privacy - Volume 1: ICISSP, ISBN 978-989-758-359-9, pages 146-154. DOI: 10.5220/0007394201460154


in Bibtex Style

@conference{icissp19,
author={José Torres and Sergio Santos and Efthimios Alepis and Constantinos Patsakis},
title={Behavioral Biometric Authentication in Android Unlock Patterns through Machine Learning},
booktitle={Proceedings of the 5th International Conference on Information Systems Security and Privacy - Volume 1: ICISSP,},
year={2019},
pages={146-154},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007394201460154},
isbn={978-989-758-359-9},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 5th International Conference on Information Systems Security and Privacy - Volume 1: ICISSP,
TI - Behavioral Biometric Authentication in Android Unlock Patterns through Machine Learning
SN - 978-989-758-359-9
AU - Torres J.
AU - Santos S.
AU - Alepis E.
AU - Patsakis C.
PY - 2019
SP - 146
EP - 154
DO - 10.5220/0007394201460154