Optimized Predictive Models in Health Care Using Machine Learning

12Wearable Gait Authentication: A Framework for Secure User Identification in Healthcare

Swathi A.¹, Swathi V.¹, Shilpa Choudhary²* and Munish Kumar³

¹Sreyas Institute of Engineering and Technology, Hyderabad, India

²CSE (AIML) Department, Neil Gogte Institute of Technology, Hyderabad, India

³Department of E&CE, Maa Saraswati Institute of Engineering and Technology, Kalanaur, India

Abstract

With the wide use of wearable Internet of Things (IoT) devices, it is now possible to quickly gather information on human activities, including unconscious or subconscious acts. One such behavior is walking, which can provide individual patterns for each person and be utilized as a biometric feature for user authentication in healthcare systems. This paper proposes the lightweight seamless authentication framework (LiSA-G), which may be used to identify and authenticate users on commercial smart watches by leveraging statistical data and sensor data features related to human behavior. With a mean equal error rate (EER) of 8.2% and fewer features and less sensor data, the study’s findings demonstrate that this approach delivers higher authentication accuracy. Such a strategy is also more realistic and can be deployed quickly due to the limited computational and energy capacity of wearable IoT systems in healthcare setting.

Keywords: Gait recognition, gait authentication

12.1 Introduction

The authentication is the process to verify the user is genuine or not. There are different processes ...