ABSTRACT
The year 2019 saw an unusual epidemic known as COVID-19, which affected the entire world. COVID-19 is a kind of coronavirus that causes widespread respiratory system damage and severe respiratory symptoms, which are associated with increased ICU admissions and mortality. The lack of therapy has prompted study in various sectors to address it. Contributions in Computer Science mostly involve the creation of algorithms for the detection, diagnosis, and prediction of COVID-19 instances.The most extensively utilized approaches in this field are machine learning (ML) and deep learning (DL). As a result, the purpose of this work was to develop a method for early detection of COVID-19 using chest X-rays images and CT scans images using different Artificial Intelligence (AI) techniques. These images were identified utilizingseveral AI algorithms, and their performance was then analyzed in order to identify the best of them. Convolutional neural networks (CNN), K nearest neighbor (KNN), Random Forest (RF), and support vector machine (SVM) among the techniques used.CNN is used in two scenarios here: the first to categorize X-ray images and CTscansimages using a softmax (with X-Ray dataset) and sigmoid (with CT dataset) classifier, and the second to extract automatic features from the images and send them to other classifiers (SVM, KNN, and RF). Given to the results, numerous classifiers work well, with most achieving over 99% accuracy. The highest accuracy was achieved in this study when using the hybrid system CNN-SVM with X-Ray database, where the test accuracy rate was 99.51%.