Multi-modal authentication system for smartphones using face, iris and periocular

Abstract

Secure authentication for smartphones is becoming important for many applications such as financial transactions. Until today PIN and password authentication are the most commonly used methods for smartphone access control. Specifically for a PIN and limited length passwords, the level of security is low and thus can be compromised easily. In this work, we propose a multi-modal biometric system, which uses face, periocular and iris biometric characteristics for authentication. The proposed system is tested on two different devices - Samsung Galaxy S5 smartphone and Samsung Galaxy Note 10.1 tablet. An extensive set of experiments conducted using the proposed system shows the applicability for secure authentication scenarios. The proposed system is tested using uni-modal and multi-modal approach. An Equal Error Rate (EER) of 0.68% is obtained from the experiments validating the robust performance of the proposed system.