Mountain Top University Library

ABSTRACT
Coronavirus disease 19 (COVID-19) is a pathogenic and highly transmissible virus infection,
where the letters CO, VI, and D stand for corona, virus, and disease, respectively. This disease is
caused by the severe acute respiratory syndrome coronavirus (SARS-CoV-2) that resulted in a
global pandemic and a significant loss of human life (Salvia, et al., 2021). This research explains
and identifies the use of deep learning techniques (residual networks) to diagnose COVID-19 from
lung ultrasound imagery. It identifies the advantages of deep learning with medical images and
gives reasons why lung ultrasound is considered a viable tool for covid-19 diagnosis. Multiple
journals, articles, reports...etc. on works related to deep learning for covid-19 diagnosis using lung
ultrasounds were reviewed. The model was developed using the POCUS dataset and for
performance analysis was compared to our baseline model (POCOVID-NET). Finally, it concludes
that diagnosis of COVID-19 could be aided by deep learning approaches for computer-assisted
interpretation of lung ultrasound imagery and affirms that RESNET-18 can be used to build a
viable computer-assisted diagnosis method for COVID-19.