PASSWORD PROTECTION APPLICATION

Abstract

This study presents a novel approach to enhancing device security for touch-based interfaces, applicable to mobile phones (iOS and Android) and Automated Teller Machines (ATMs). Our approach integrates keystroke dynamics and mobile sensors into user authentication through an innovative algorithmic program. By leveraging keystroke dynamics as biometric authentication, focusing on typing speed and sensor data, we address inherent vulnerabilities in traditional password systems that are increasingly targeted in today's digital environment. The proposed algorithm continuously learns from the unique timing patterns of each user's keystrokes, resulting in a personalized and adaptive security system. We evaluated this timing-centric method on various smartphone models and considered its implications for ATM security. The results, based on keystroke timing accuracy and sensor data analysis, show a significant enhancement in authentication accuracy, underscoring the practical value of our approach in fortifying defenses against prevalent cyber threats. This paper details the machine learning techniques used, assesses system effectiveness, and provides comprehensive insights into improving mobile security programs. Our findings contribute to addressing growing security concerns for touch-based devices, particularly within smartphone operating systems.

Date of Award	2024
Original language	American English
Awarding Institution	HBKU College of Science and Engineering