Modified Convolutional Network for the Identification of Covid-19 with a Mobile System

Abstract

In this paper, we modified a low-cost and rapid method to detect chest X-rays based on MobileNet. Because MobileNet is a lightweight neural network, we modified and optimized backpropagation learning to train the model. In the subsequent COVID-19, pneumonia, and normal tests, the recognition accuracy reached 99.14%, which greatly improved the performance of the model. Our scheme can produce an effective model suitable for low-performance mobile devices.