The role of pre-trained models in diagnosing Covid-19 using chest X-ray images

Ba Alawi, A. E.; Saeed, A. Y. A.; Rassam, M. A.

The recent outbreak of the novel coronavirus affects the globe. It was in a major city of China called 'Wuhan' where COVID-19 first appeared. This new disease is caused by a novel, or new coronavirus and deemed to be a worldwide pandemic. This extreme virus, which spreads by human contact, is now invading more than two hundred countries across the world. In comparison, new coronavirus signs are very close to the general seasonal influenza. The screening of infected people in the war against COVID-19 is seen as a crucial step for controlling and fighting the spread of this infectious virus. Rapid diagnosis is of so high priority that suspected cases are identified as early as possible to guarantee a control over the spread of this disease. This paper investigates the use of pre-trained models in diagnosing Covid-19 infected patients using Chest X-ray (CXR). The transfer learning techniques have been applied to pre-trained models for an image classification task. The result obtained showed that these models have a high performance in the diagnosis process. Results are so promising that they reached more than 94% in testing cases. © 2021 IEEE.