COVID-19 Detection Using Deep Learning Methods

Abstract

According to the world health organization, the coronavirus epidemic threatens the world's health system every day. Health resources in most countries are either insufficient or not fairly shared. There are various problems such as the number of health personnel, the number of beds, or the number of intensive care units. Using limited resources at the optimum level is the key to the country's health systems to overcome this epidemic. Disease detection is an important factor in preventing the epidemic. The higher the success, the more controlled the spread of the virus. Whether the person has a virus or not is usually done by the PCR test. In addition to the PCR method, chest X-ray images can be classified with deep learning methods. Deep learning methods have become popular in academic studies by processing multi-layered images in one go and by defining manually entered parameters in machine learning. This popularity reflected positively on limited health datasets. In this study, it was aimed to detect the disease of people whose X-rays were taken for suspected COVID-19. In such COVID-19 studies, a binary classification has generally been made. The data set includes chest X-rays of patients with COVID-19, viral pneumonia, and healthy patients. Before the classification process, the data augmentation method was applied to the data set. These three groups have been classified through multi-class classification deep learning models. © 2020 IEEE.