COVID-Net UV: An End-to-End Spatio-Temporal Deep Neural Network Architecture for Automated Diagnosis of COVID-19 Infection from Ultrasound Videos

Hilda Azimi; Ashkan Ebadi; Jessy Song; Xi Pengcheng; Stéphane Tremblay; Alexander Wong

doi:10.15353/jcvis.v8i1.5359

Vol. 8 No. 1 (2022)
Special issue: Proceedings of CVIS 2022

Articles

COVID-Net UV: An End-to-End Spatio-Temporal Deep Neural Network Architecture for Automated Diagnosis of COVID-19 Infection from Ultrasound Videos

https://doi.org/10.15353/jcvis.v8i1.5359

Published 2023-05-10

Hilda Azimi
Ashkan Ebadi
Jessy Song
Xi Pengcheng
Stéphane Tremblay
Alexander Wong

How to Cite

Azimi, H., Ebadi, A., Song, J., Pengcheng, X., Tremblay, S., & Wong, A. (2023). COVID-Net UV: An End-to-End Spatio-Temporal Deep Neural Network Architecture for Automated Diagnosis of COVID-19 Infection from Ultrasound Videos. Journal of Computational Vision and Imaging Systems, 8(1), 11–13. https://doi.org/10.15353/jcvis.v8i1.5359

Download Citation

Abstract

Besides vaccination, as an effective way to mitigate the further spread of COVID-19, fast and accurate screening of individuals to test for the disease is yet necessary to ensure public health safety. We propose COVID-Net UV, an end-to-end hybrid spatio-temporal deep neural network architecture, to detect COVID-19 infection from lung point-of-care ultrasound videos captured by convex transducers. The COVID-Net UV comprises a convolutional neural network that extracts spatial features and a recurrent neural network that learns temporal dependence. After careful hyperparameter tuning, the network achieves an average accuracy of 94.44\% with no false-negatives for COVID-19 cases. The goal of COVID-Net UV is to assist front-line clinicians in the fight against COVID-19 as a decision support tool via accelerating the screening of lung point-of-care ultrasound videos and automatic detection of COVID-19 positive cases.

pdf