Research

[Medrxiv] Online COVID-19 diagnosis with chest CT images: Lesion-attention deep neural networks (preprint)

Medrxiv, 29 May 2020

(This article is a preprint and has not been certified by peer review)

Bin Liu, Xiaoxue Gao, Mengshuang He, Fengmao Lv, Guosheng Yin

Highlights:

To help alleviate burdens on medical systems, researchers developed a lesion-attention deep neural network (LA-DNN) to predict COVID-19 positive or negative with a richly annotated chest CT image dataset. The proposed data-efficient LA-DNN model focuses on the primary task of binary classification for COVID-19 diagnosis, while an auxiliary multi-label learning task is implemented simultaneously to draw the model’s attention to the five lesions associated with COVID-19. The joint task learning process makes it a highly sample-efficient deep neural network that can learn COVID-19 radiology features more effectively with limited but high-quality, rich-information samples. Researchers said that the experimental results reached the clinical standards for practical use.

Key features include:

(1) Fast: Require only chest CT image to show the diagnostic result immediately

(2) Accurate: AUC: 93%, Accuracy: 88%, Sensitivity: 86%, Specificity: 90%

(3) Easy to Use: Online web with user-friendly interface (https://www.covidct.cn/)

(4) Open Source: All codes and data are freely accessible at github
(https://github.com/xiaoxuegao499/LA-DNN-for-COVID-19-diagnosis)

Read full text

Tags
Show More

Related Articles

Back to top button
Close