Simple shape-based touch behavioral biometrics authentication for smart mobiles

Abstract

One of the main concerns during usage of the current smart mobile devices in public is the vulnerability of password hacking by shoulder-suffering or smudge attack. The traditional user authentication techniques such as PIN code or patterns-based password are easy target of such attacks. In this paper, we propose using simple shapes (e.g., circle, triangle, etc.) to get users' touch behavioral biometrics data. The users are asked to draw over these shapes while the developed system extracts 25 different features (e.g., finger middle stroke and its pressure, velocity, mobile orientation, etc.) for the model training and authentication purpose. The proposed solution is simple for all kinds of users and could solve the problem of password hacking through shoulder-suffering or smudge attacks.