Automatic detection of covid-19 pneumonia in chest computed tomography scans using convolutional neural networks

Abstract

The Coronavirus outbreak caused by the SARSCoV-2 virus has been the focal point of global attention over the past two years, owing to the pandemic’s infection rate and the huge burden on the world’s healthcare systems and economy. Diagnosis of infection by the virus may be carried out through a number of tests, with the current mainly used technique being reverse transcription polymerase chain reaction tests. An alternative approach for diagnosis is through the use of medical imagery such as chest X-rays or chest Computed Tomography images. In this work, we propose a machine learning driven approach which automatically detects pulmonary pathological features caused by the Coronavirus infection in chest Computed Tomography images. The model was trained and evaluated on the COVIDx CT-2A dataset, achieving an accuracy of 96.31% on the testing segment of the dataset.