Behavioral Biometrics Scheme with Keystroke and Swipe Dynamics for User Authentication on Mobile Platform

Abstract

Due to the explosive growth of mobile devices worldwide, authentication is receiving increasing attention. Conventionally, explicit authentication methods such as password is employed. However, the system would be breached if the password is stolen. Therefore, there is a continual search for ways to strengthen authentication for mobile platforms. Behavioral biometric information such as keystroke and swipe dynamics can be used for enhancing security. This paper presents an authentication scheme which employs a combination of password, keystroke dynamics, and swipe dynamics for touchscreen mobile devices. Features extracted from swiping pattern and typing pattern were evaluated. Accuracy of the system was enhanced by using combined behavioral biometrics features, as compared with using only a single set of features. The identification accuracies increased significantly from the range of 63.03% - 88.30% to 86.59% -94.05%; while the F1 scores increased from the range of 60.42% - 85.96% to 85.43% - 93.15%.