Abstract
Novel Coronavirus disease (COVID-19) is a highly contagious respiratory infection that has had devastating effects on the world. Recently, new COVID-19 variants are emerging making the situation more challenging and threatening. Evaluation and quantification of COVID-19 lung abnormalities based on chest Computed Tomography (CT) images can help determining the disease stage, efficiently allocating limited healthcare resources, and making informed treatment decisions. During pandemic era, however, visual assessment and quantification of COVID-19 lung lesions by expert radiologists become expensive and prone to error, which raises an urgent quest to develop practical autonomous solutions. In this context, first, the paper introduces an open-access COVID-19 CT segmentation dataset containing 433 CT images from 82 patients that have been annotated by an expert radiologist. Second, a Deep Neural Network (DNN)-based framework is proposed, referred to as the text {COVID-Rate}, that autonomously segments lung abnormalities associated with COVID-19 from chest CT images. Performance of the proposed text {COVID-Rate} framework is evaluated through several experiments based on the introduced and external datasets. Third, an unsupervised enhancement approach is introduced that can reduce the gap between the training set and test set and improve the model generalization. The enhanced results show a dice score of 0.8069 and specificity and sensitivity of 0.9969 and 0.8354, respectively. Furthermore, the results indicate that the text {COVID-Rate} model can efficiently segment COVID-19 lesions in both 2D CT images and whole lung volumes. Results on the external dataset illustrate generalization capabilities of the text {COVID-Rate} model to CT images obtained from a different scanner.
Highlights
Coronavirus disease 2019 (COVID-19) has been the world’s most threatening challenge of the twenty first century
We propose the COVID-Rate framework, which is a Deep Learning (DL)-based model for segmenting COVID-19 lesions, including Ground-Glass Opacity (GGO) and consolidations
Ref.[25] implemented a specific data augmentation approach into a generative adversarial segmentation network and achieved improved results. Capitalizing on this vision, we propose a specific approach for generating synthetic pairs of Computed Tomography (CT) images and their corresponding infection masks by extracting the COVID-19 regions of infection from infected chest CT images and inserting them into healthy chest CT images
Summary
Coronavirus disease 2019 (COVID-19) has been the world’s most threatening challenge of the twenty first century. Evaluation and quantification of lung involvement in COVID-19 patients based on their chest images can help determine the disease stage, have an optimal allocation of the limited health resources, and make informed treatment decisions. Radiologists measure the COVID-19 lesions from the chest CT images and quantify the disease’s severity using different severity measures such as the Percentage of Opacity (PO) and CT severity score. The PO indicates the extent of involvement of the whole lung v olume[14], while the CT severity score is determined based on the spread of the COVID-19 lesions in each lobe[15]. Automatic segmentation of infectious regions can, help quantify the extent of lung involvement in patients confirmed with COVID-19, compute different severity scores, and speed up the treatment procedure. The main contributions of this study are as follows:
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
