Deep learning-based feature extraction coupled with multi class SVM for COVID-19 detection in the IoT era

The deadly coronavirus virus (COVID-19) was confirmed as a pandemic by the World Health Organisation (WHO) in December 2019. Prompt and early identification of suspected patients is necessary to monitor the transmission of the disease, increase the effectiveness of medical treatment and as a result, decrease the mortality rate. The adopted method to identify COVID-19 is the Reverse-Transcription Polymerase Chain Reaction (RT-PCR), the method is affected by the shortage of RT-PCR kits and complexity. Medical imaging using deep learning has proved to be one of the most efficient methods of detecting respiratory diseases, but efficient deep learning architecture and low data are affecting the performance of the deep learning models. To detect COVID-19 efficiently, a deep learning model based feature extraction coupled with support vector machine (SVM) was employed in this study, Seven pre-trained models were employed as feature extractors and the extracted features are classified by multi-class SVM classifier to classify COVID-19, common pneumonia and healthy individuals' CT scan images, to improve the performance of the models and prevent overfitting, training was also carried out on augmented images. To generalise the model's performance and robustness, three datasets were merged in the study. The model with the best performance is the VGG19 which was trained with augmented images, the VGG19 achieved an accuracy of 96%, sensitivity of 0.936, specificity of 0.967, F1 score of 0.935, precision of 0.934, Yonden Index of 0.903 and AUC of 0.952. The best model shows that COVID-19 can be detected efficiently on CT scan images.