In recent years, with the widespread adoption of smartphones, security measures have become increasingly important. Multifactor authentication is one method used to strengthen authentication. This paper explores the application of multifactor authentication for smartphones. Specifically, we consider a method that combines 1) possession of a smartphone (possession-based information), 2) knowledge of a password or pattern (knowledge-based information), and 3) biometric authentication, such as fingerprint, face, or voice recognition. Such a smartphone may already be in someone else's possession when unauthorized access is attempted, it is crucial to combine these factors. This paper proposes an enhancement to the screen unlocking process through improved pattern authentication. Pattern authentication typically involves registering a sequence of nine points on the device, which the user must trace in the correct order to unlock it. However, since this method relies solely on the trace pattern for authentication, it is vulnerable if a third party memorizes the pattern. To address this issue, we aim to improve the security of pattern authentication by developing a machine learning model that captures the unique behavioral traits of the user. This model analyzes not only the passing trajectory of the points but also the entirety of the user's interaction with the device as a dataset.

Author: Ayahiko Niimi

Published in: International Conference for Internet Technology and Secured Transactions (ICITST-2024)

• Date of Conference: 4-6 November 2024
• DOI: 10.20533/ICITST.2024.0023
• ISBN: 978-1-913572-76-1
• Conference Location: St Anne’s College, Oxford University, UK