Comparative Study of Deep Learning Models for COVID-19 Diagnosis

ABSTRACT

Corona Virus Disease 2019(COVID-19) spread far and wide in numerous nations in early 2020, causing the world to face an existential health crisis. This pandemic continues to have a devastating effect on the global population and by now it has infected more than a few million individuals around the world. One significant obstacle in controlling the spreading of this virus is that the initial system for addressing this infectious disease was not clear. A basic advancement in the struggle opposite the COVID-19 pandemic is early screening and dependable diagnosis utilizing computerized detection of lung infections. Computed Tomography (CT) scans and X-rays imagery offers great potential help to clinical specialists tackling COVID-19. An efficient Deep Learning diagnosis application needs to be developed so that accurate and precise prediction can be done for the disease. This paper introduces dataset analysis and comparative evaluation of deep learning models for creating disease diagnosis application using image processing. Comparison is done using three main deep learning models-Convolutional Neural Network (CNN), Support Vector Machine (SVM) Logistic Regression. Dataset analysis and model selection is a crucial phase for developing a predictive deep learning algorithm. This analysis is done for better results and is done using Orange data mining software. © 2021 IEEE.