COVID-19 Detection from X-ray Images Using Different Artificial Intelligence Hybrid Models
Ali Mohammad Alqudah, Shoroq Qazan, Hiam Alquran, Isam Abu Qasmieh, Amin Alqudah |Pages: 168-178|

Abstract—COVID-19 leads to severe respiratory symptoms that are associated with highly intensive care unit (ICU) admissions and deaths. Early diagnosis of coronavirus limits its wide spread. Real-time reverse transcription-polymerase chain reaction (RT-PCR) is the strategy that has been used by clinicians to discover the presence or absence of this type of virus. This technique has a relatively low positive rate in the early stage of this disease. Therefore, clinicians call for other ways to help in the diagnosis of COVID-19. The appearance of X-ray chest images in case of COVID-19 is different from any other type of pneumonic disease. Therefore, this research is devoted to employ artificial intelligence techniques in the early detection stages of COVID-19 from chest X-ray images. Different hybrid models – each consists of deep features extraction and classification techniques – are implemented to assist clinicians in the detection of COVID-19. Convolutional neural network (CNN) is used to extract the graphical features in the implementations of the hybrid models from the chest X-ray images. The classification, to COVID-19 or Non-COVID-19, is achieved using different machine learning algorithms such as CNN, support vector machine (SVM), and random forest (RF), to obtain the best recognition performance. The most significant two extracted features are employed for training and parameters testing. According to the performance results of the designed models, CNN outperforms other classifiers with a testing accuracy of 95.2%.