Clinical Sound Dataset for Covid-19 and Evaluation using ResNet

In this study, we built a sound dataset for COVID-19. Unlike other studies, we collect clinical data from patients with COVID-19 who agree to the Informed Consent Form. Therefore, the sound data from patients with COVID-19 are from subjects who were diagnosed with COVID-19 at the hospital. The dataset consists of 600 sounds which are 300 from patients with COVID-19 and 300 from non-COVID-19 subjects. For validation, we use ResNet to classify COVID-19 and non-COVID-19 sounds.