DenseCapsNet: Detection of COVID-19 from X-ray images using a capsule neural network. (June 2021)

• Record Type: Journal Article
• Authors: Quan, Hao; Xu, Xiaosong; Zheng, Tingting; Li, Zhi; Zhao, Mingfang; Cui, Xiaoyu
• Is Part Of: Computers in biology and medicine. Volume 133(2021)
• Publication Date: 2021-06
• Subjects: COVID-19 -- Chest X-ray -- Classification -- Deep learning -- Capsule neural network; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2021.104399
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: At present, the global pandemic as it relates to novel coronavirus pneumonia is still a very difficult situation. Due to the recent outbreak of novel coronavirus pneumonia, novel chest X-ray (CXR) images that can be used for deep learning analysis are very rare. To solve this problem, we propose a deep learning framework that integrates a convolutional neural network and a capsule network. DenseCapsNet, a new deep learning framework, is formed by the fusion of a dense convolutional network (DenseNet) and the capsule neural network (CapsNet), leveraging their respective advantages and reducing the dependence of convolutional neural networks on a large amount of data. Using 750 CXR images of lungs of healthy patients as well as those of patients with other pneumonia and novel coronavirus pneumonia, the method can obtain an accuracy of 90.7% and an F1 score of 90.9%, and the sensitivity for detecting COVID-19 can reach 96%. These results show that the deep fusion neural network DenseCapsNet has good performance in novel coronavirus pneumonia CXR radiography detection. Highlights: A preprocessing method is proposed to alleviate the problem of image heterogeneity. Based on DenseNet and CapsNet, the deep learning framework DenseCapsNet is proposed. The sensitivity of COVID-19 based on DenseCapsNet was 96%. Location of COVID-19 lesions.