COVID-19 Diagnosis from CT Imaging using Imaging and Machine Analysis

Abstract

Coronavirus (COVID-19) is spreading rapidly around the world and, as of October 2020, more than 1,966,000 people have been infected in more than 200 countries. Early detection of COVID-19 is essential for the provision and protection of HIV-negative people in adequate health care for patients. To do this, we developed an automated diagnostic program for COVID-19 from pneumonia (CPA) obtained from chest tomography (CT). We propose, in particular, the Noise Resilient method of machine learning that focuses on regions of lung infection while making diagnostic decisions. Note that the sizes of the infection sites between COVID-19 and CAP are not well measured, in part due to the rapid progression of COVID-19 after the onset of symptoms. Large amounts of CVID-19 CT data from hospitals have been used to evaluate our frameworks.