SUMMARY
Coronavirus disease (COVID-19) is a pandemic disease that has spread rapidly among people living in many countries. The effective screening and immediate medical response for the infected patients are important to treat with stopping the spread of COVID-19 disease. Chest radiography (CXR) image is usually required for lung severity assessment. However, chest X-rays in COVID-19 interpretation is required expert radiologists’ knowledge. This study aims to improve the COVID-19 X-ray image classification by feature selection technique using the regression mutual information deep convolution neuron networks (RMI Deep-CNNs). The dataset consists of 219 COVID-19, 500 viral pneumonias, and 500 normal chest X-ray images. CXR images were comprehensively pre-trained using DCNNs to extract image features, then, the critical features were selected using regression mutual information followed by the fully connected with softmax layer for classification. These networks were compared for the classification of two different schemes (ResNet152V2 and InceptionV3). The classification accuracy, sensitivity, and specificity for both schemes were 92.21%, 100%, 90% and 91.39%, 100%, 82.50%, respectively. In addition, RMI Deep-CNNs not only improve the accuracy but also reduce trainable features by over 80%. This approach tends to significantly improve the computation time and model accuracy for COVID-19 classification.