Skip to main content

Deep learning classification of COVID-19 in chest radiographs: performance and influence of supplemental training.

Publication ,  Journal Article
Fricks, RB; Ria, F; Chalian, H; Khoshpouri, P; Abadi, E; Bianchi, L; Segars, WP; Samei, E
Published in: J Med Imaging (Bellingham)
November 2021

Purpose: Accurate classification of COVID-19 in chest radiographs is invaluable to hard-hit pandemic hot spots. Transfer learning techniques for images using well-known convolutional neural networks show promise in addressing this problem. These methods can significantly benefit from supplemental training on similar conditions, considering that there currently exists no widely available chest x-ray dataset on COVID-19. We evaluate whether targeted pretraining for similar tasks in radiography labeling improves classification performance in a sample radiograph dataset containing COVID-19 cases. Approach: We train a DenseNet121 to classify chest radiographs through six training schemes. Each training scheme is designed to incorporate cases from established datasets for general findings in chest radiography (CXR) and pneumonia, with a control scheme with no pretraining. The resulting six permutations are then trained and evaluated on a dataset of 1060 radiographs collected from 475 patients after March 2020, containing 801 images of laboratory-confirmed COVID-19 cases. Results: Sequential training phases yielded substantial improvement in classification accuracy compared to a baseline of standard transfer learning with ImageNet parameters. The test set area under the receiver operating characteristic curve for COVID-19 classification improved from 0.757 in the control to 0.857 for the optimal training scheme in the available images. Conclusions: We achieve COVID-19 classification accuracies comparable to previous benchmarks of pneumonia classification. Deliberate sequential training, rather than pooling datasets, is critical in training effective COVID-19 classifiers within the limitations of early datasets. These findings bring clinical-grade classification through CXR within reach for more regions impacted by COVID-19.

Duke Scholars

Altmetric Attention Stats
Dimensions Citation Stats

Published In

J Med Imaging (Bellingham)

DOI

ISSN

2329-4302

Publication Date

November 2021

Volume

8

Issue

6

Start / End Page

064501

Location

United States

Related Subject Headings

  • 4003 Biomedical engineering
  • 3202 Clinical sciences
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Fricks, R. B., Ria, F., Chalian, H., Khoshpouri, P., Abadi, E., Bianchi, L., … Samei, E. (2021). Deep learning classification of COVID-19 in chest radiographs: performance and influence of supplemental training. J Med Imaging (Bellingham), 8(6), 064501. https://doi.org/10.1117/1.JMI.8.6.064501
Fricks, Rafael B., Francesco Ria, Hamid Chalian, Pegah Khoshpouri, Ehsan Abadi, Lorenzo Bianchi, William P. Segars, and Ehsan Samei. “Deep learning classification of COVID-19 in chest radiographs: performance and influence of supplemental training.J Med Imaging (Bellingham) 8, no. 6 (November 2021): 064501. https://doi.org/10.1117/1.JMI.8.6.064501.
Fricks RB, Ria F, Chalian H, Khoshpouri P, Abadi E, Bianchi L, et al. Deep learning classification of COVID-19 in chest radiographs: performance and influence of supplemental training. J Med Imaging (Bellingham). 2021 Nov;8(6):064501.
Fricks, Rafael B., et al. “Deep learning classification of COVID-19 in chest radiographs: performance and influence of supplemental training.J Med Imaging (Bellingham), vol. 8, no. 6, Nov. 2021, p. 064501. Pubmed, doi:10.1117/1.JMI.8.6.064501.
Fricks RB, Ria F, Chalian H, Khoshpouri P, Abadi E, Bianchi L, Segars WP, Samei E. Deep learning classification of COVID-19 in chest radiographs: performance and influence of supplemental training. J Med Imaging (Bellingham). 2021 Nov;8(6):064501.

Published In

J Med Imaging (Bellingham)

DOI

ISSN

2329-4302

Publication Date

November 2021

Volume

8

Issue

6

Start / End Page

064501

Location

United States

Related Subject Headings

  • 4003 Biomedical engineering
  • 3202 Clinical sciences