Chest X-Ray Imaging Severity Score of COVID-19 Pneumonia

Abstract

Despite the decrease in COVID-19 cases worldwide due to the development of extensive vaccination campaigns and effective containment measures adopted by most countries, this disease continues to be a global concern. Therefore, it is necessary to continue developing methods and algorithms that facilitate decision-making for better treatments. This work proposes a method to evaluate the degree of severity of the affectations caused by COVID-19 in the pulmonary region in chest X-ray images. The proposed algorithm addresses the problem of confusion between pulmonary lesions and anatomical structure (i.e., bones) in chest radiographs. In this paper, we adopt the Semantic Genesis approach for classifying image patches of the lung region into two classes (affected and unaffected). Experiments on a database consisting of X-rays of healthy people and patients with COVID-19 have shown that the proposed approach provides a better assessment of the degree of severity caused by the disease.