COVID-19 Diagnosis from Chest X-Ray Images Using Machine Learning Approach

During the COVID-19 outbreak which has dramatic consequences that affected people around the world, many of the governments in cooperation with the health care institutes were seeking to find out a low-cost, easy, fast, and accurate COVID-19 detection method to control the spreading of this infectious disease. The dominant testing method that is widely used for COVID-19 diagnosis is the Polymerase Chain Reaction (PCR). However, the PCR testing method is time-consuming and has a false negative probability that can't be neglected. This paper proposed a machine learning (ML) model for COVID-19 detection from Chest X-Ray (CXR) images. The developed system extracted a group of first and second order statistical features and only the most significant features were selected by using the t-test. The proposed model was trained using 4500 CXR images that consist of 2250 of normal and 2250 of COVID-19 samples. The testing dataset comprises of the same number of normal and COVID-19 cases and the total number of testing set is equal to 1500 CXR images. The proposed model was able to effectively diagnose all the testing samples with an accuracy of 100%.