Smart Pre-Screening of Patients for COVID-19 Using Depth-Wise Separableconvolutional Neural Network

Abstract

Hospitals have become overcrowded and overburdened and at the same time have become a major source for viruses and pathogens to breed and spread. Even the testing mechanism and kits are limited in number and are very expensive to manufacture. We propose a mechanism using Machine Learning, Neural network techniques, and hardware to create a commercial inexpensive pre-screening device for such diseases. This device is designed for conditions that can be easily detected by analyzing the cough audio signals and have distinct cepstral coefficients. This device can then help us reduce the burden on the healthcare system as we can pre-screen and only send patients that are more likely infected to the hospital.