Abstract
SARS-CoV-2 is a severe respiratory syndrome-associated coronavirus that triggered the global pandemic of coronavirus disease 2019, claiming millions of lives. However, the ongoing spread of SARS-CoV-2 poses a great diagnostic challenge, as the symptoms of Covid-19 are often difficult to distinguish from those of other respiratory diseases. To meet this challenge, we urgently need a timely and cost-effective method to detect affected patients. In this study, we adopt X-ray images to identify Covid-19 patients, utilize a dataset of 3000 images containing Covid-19 positive and normal samples, and give full play to the excellent performance of transfer learning. Using a model ensemble approach, we were able to accurately identify coronavirus-infected patients from chest X-ray images. The proposed ensemble model achieves 97.74% and 99.67% in terms of ACC and AUC, respectively. We look forward to this excellent research to provide strong support for medical staff in diagnosing other respiratory diseases similar to Covid-19.
Funding Information
  • the Medico-Engineering Cooperation Funds from university of Electronic Science and Technology of China (No. ZYGX2021YGLH213, No. ZYGX2022YGRH016)
  • the Municipal Government of Quzhou (Grant 2022D018, Grant 2022D029)