Diagnosis of COVID-19 from Chest CT Images Based on Deep Learning

Novel coronavirus disease (COVID-19) is an extremely contagious and quickly spreading coronavirus infestation, it was recognized in December 2019 and has posed critical challenges for the public health, research, and medical communities. In the fight against this novel disease, there is a pressing need for rapid and effective screening tools to identify patients infected with COVID-19, and to this end Computed Tomography imaging has been proposed as one of the key screening methods. In this paper, CT image dataset is preprocessed to get probable COVID-19 CT images which are associated with COVID-19 pneumonia (so-called prob-CT). Based on it, a diagnostic method for CT chest images using deep Convolutional Neural Networks (CNN) models is presented. Here the data includes hundreds of images obtained by one computed tomography and is fed into the proposed model for deep feature extractions. We analyse and evaluate the performance of the model in COVID-19 diagnosis. The result obtained by semiquantitative analysis of preferred measures shows that the proposed model performs well with accuracy (92.1%) and sensitivity (93.6%).