CGENet: A Deep Graph Model for COVID-19 Detection Based on Chest CT

SIMPLE SUMMARY: This study proposes a new COVID-19 detection system called CGENet, based on computer vision and chest computed tomography images. First, an optimal backbone selection algorithm was proposed to determine the best backbone network for the CGENet adaptively. Then, we introduced a novel graph embedding mechanism to fuse the spatial relationship into the feature vectors. Finally, we chose the extreme learning machine as the classifier of the proposed CGENet to boost the classification performance. The proposed CGENet was evaluated on a public dataset using 5-fold cross-validation and compared with other algorithms. The results revealed that the proposed model achieved state-of-the-art classification performance. In all, the CGENet can be an effective and efficient tool that can assist COVID-19 diagnosis. ABSTRACT: Accurate and timely diagnosis of COVID-19 is indispensable to control its spread. This study proposes a novel explainable COVID-19 diagnosis system called CGENet based on graph embedding and an extreme learning machine for chest CT images. We put forward an optimal backbone selection algorithm to select the best backbone for the CGENet based on transfer learning. Then, we introduced graph theory into the ResNet-18 based on the k-nearest neighbors. Finally, an extreme learning machine was trained as the classifier of the CGENet. The proposed CGENet was evaluated on a large publicly-available COVID-19 dataset and produced an average accuracy of 97.78% based on 5-fold cross-validation. In addition, we utilized the Grad-CAM maps to present a visual explanation of the CGENet based on COVID-19 samples. In all, the proposed CGENet can be an effective and efficient tool to assist COVID-19 diagnosis.