Joint Use of Vital Signs and Cough Sounds for Pandemic Detection

Abstract

In response to the challenges once posed by the COVID-19 pandemic, this paper presents a comprehensive solution that integrates advanced techniques to enhance the detection of infections remotely, using sensors on a wearable bracelet. Building on our previous work, we introduce a Machine Learning model that can classify COVID-19 and Healthy patients from their cough sounds and vital signs. Health data like Body Temperature, Heart Rate, and SpO2 levels are collected by a sensor in the wristband and are sent to the mobile application for diagnosis. The system is connected to a local backend server that performs the classification process. The results from the Cough Classification and Vital Signs classification contribute to a robust assessment of infection probability. Our results show a significant improvement in detection accuracy, indicating the potential of this solution to serve as an adaptable tool in future pandemics with respiratory symptoms.