Multi-modal Biometric Authentication: Harnessing Human Gait and Keystroke Dynamics for Enhanced Security

Multi-modal Biometric Authentication: Harnessing Human Gait and Keystroke Dynamics for Enhanced Security

연구 분야: Analysis

학회: International Conference on Computational Technologies and Electronics

In an era of escalating cyber threats, user authentication has become paramount in safeguarding sensitive information. Traditional methods like passwords and PINs are proving inadequate in the face of sophisticated attacks. This article proposes a cutting-edge approach that leverages the unique biometric signatures of human gait and keystroke dynamics for robust user authentication. Through a combination of sensor technologies and advanced machine learning algorithms, this method offers a multi-layered security solution that promises to revolutionize the authentication landscape. A group of fifty-two volunteers with varying ages, professions, and levels of education was chosen for the preparation of the study database. The keystroke pattern and gait of each individual were captured utilizing built-in smartphone sensors in four different sessions. To capture human keystroke patterns four different fixed-texts are considered. Afterward, both keystroke and gait patterns underwent comprehensive analysis employing multiple machine learning algorithms, both in isolation and in combination. The findings indicate that the concurrent utilization of keystroke and gait patterns resulted in a markedly elevated degree of precision in user identification and authentication when compared to the individual application of either of these patterns. The research pathways for the development and integration of this dual-factor authentication system have been outlined properly. It is expected that this comprehensive and multi-tiered authentication approach will substantially enhance security in both physical and digital spaces while ensuring a layer of convenience for the users.