PERFORMANCE TESTING OF COVID-19 IMAGE CLASSIFICATION USING DIFFERENT MACHINE LEARNING ARCHITECTURES

Abstract

The outbreak of the Covid-19 pandemic brought the need for research in this field to assist in diagnosing patients more accurately and with more speed than is possible using just human resources. Deep Learning and especially Convolutional Neural Network has shown great accuracy levels in medical field tasks such as image analysis, segmentation and classification. In this research we will review several deep learning models and focus on finding the CNN architectures which achieve highest classification accuracy results. Different datasets will be used for training and testing with the purpose of classifying CT and X-Ray scans into Covid-19 and No Covid-19 to give a correct diagnosis.