COVID-19 Classification of CT Lung Images Using Intelligent Wolf Optimization Based Deep Convolutional Neural Network

Chest computed tomography (CT) imaging is highly reliable and practical in diagnosing and analyzing COVID-19, mainly in the infectious region rather than reverse-transcription polymerase chain reaction (RT-PCR). The intelligent wolf optimization-deep convolutional neural network (deep CNN) classifier is proposed to classify COVID from the CT images in this research. The texture features are acquired from three different regions of the CT chest images, namely the lung region, area, and contour using the texture descriptors, viz “local binary pattern (LBP), local optimal oriented pattern (LOOP)”, and ResNet-101-based features. The texture features form the input to the proposed intelligent wolf-based deep CNN classifier, which performs the COVID-19 classification. The proposed classifier achieves accuracy, sensitivity, and specificity of 85.32%, 85.74%, and 87.57%, respectively for the training–testing ratio of 80–20. The accuracy, sensitivity, and specificity of the intelligent wolf optimization-deep CNN classifier are 88.37%, 89.47%, and 91.29% respectively for the K-fold value of 10.