Deep Transfer Learning-Based Framework for COVID-19 Diagnosis Using Chest CT Scans and Clinical Information.

Abstract

The Coronavirus Disease 2019 (COVID-19) which first emerged in Wuhan, China in late December, 2019, has now spread to all the countries in the world. Conventional testing methods such as the antigen test, serology tests, and polymerase chain reaction tests are widely used. However, the test results can take anything from a few hours to a few days to reach the patient. Chest CT scan images have been used as alternatives for the detection of COVID-19 infection. Use of CT scan images alone might have limited capabilities, which calls attention to incorporating clinical features. In this paper, deep learning algorithms have been utilized to integrate the chest CT scan images obtained from patients with their clinical characteristics for fast and accurate diagnosis of COVID-19 patients. The framework uses an ANN to obtain the probability of the patient being infected with COVID-19 using their clinical information. Beyond a certain threshold, the chest CT scan of the patient is classified using a deep learning model which has been trained to classify the CT scan with 99% accuracy.