Reliable image classification based on evidence and GCN

With good classification performance, digital medical image methods based on deep learning have received more and more attention. However, the problem of these methods is that the black-box characteristic of deep learning makes it difficult to explain the model and prediction results. In order to improve the interpretability and reliability of medical image classification tasks, this paper designs a network structure that introduces eliminating uncertainty in GCN feedback by using evidence theory and subjective logic. The diagnosis experiments of COVID-19 images show that the model can effectively improve the recognition accuracy of COVID-19 patients, and has good reliability.