Transfer Learning-Based Automatic Detection of Coronavirus Disease 2019 (COVID-19) from Chest X-ray Images.
COVID-19
Convolution Neural Network
Deep Learning
Machine Learning
Transfer Learning
X-ray Images
Journal
Journal of biomedical physics & engineering
ISSN: 2251-7200
Titre abrégé: J Biomed Phys Eng
Pays: Iran
ID NLM: 101589641
Informations de publication
Date de publication:
Oct 2020
Oct 2020
Historique:
received:
02
08
2020
accepted:
08
09
2020
entrez:
2
11
2020
pubmed:
3
11
2020
medline:
3
11
2020
Statut:
epublish
Résumé
Coronavirus disease 2019 (COVID-19) is an emerging infectious disease and global health crisis. Although real-time reverse transcription polymerase chain reaction (RT-PCR) is known as the most widely laboratory method to detect the COVID-19 from respiratory specimens. It suffers from several main drawbacks such as time-consuming, high false-negative results, and limited availability. Therefore, the automatically detect of COVID-19 will be required. This study aimed to use an automated deep convolution neural network based pre-trained transfer models for detection of COVID-19 infection in chest X-rays. In a retrospective study, we have applied Visual Geometry Group (VGG)-16, VGG-19, MobileNet, and InceptionResNetV2 pre-trained models for detection COVID-19 infection from 348 chest X-ray images. Our proposed models have been trained and tested on a dataset which previously prepared. The all proposed models provide accuracy greater than 90.0%. The pre-trained MobileNet model provides the highest classification performance of automated COVID-19 classification with 99.1% accuracy in comparison with other three proposed models. The plotted area under curve (AUC) of receiver operating characteristics (ROC) of VGG16, VGG19, MobileNet, and InceptionResNetV2 models are 0.92, 0.91, 0.99, and 0.97, respectively. The all proposed models were able to perform binary classification with the accuracy more than 90.0% for COVID-19 diagnosis. Our data indicated that the MobileNet can be considered as a promising model to detect COVID-19 cases. In the future, by increasing the number of samples of COVID-19 chest X-rays to the training dataset, the accuracy and robustness of our proposed models increase further.
Sections du résumé
BACKGROUND
BACKGROUND
Coronavirus disease 2019 (COVID-19) is an emerging infectious disease and global health crisis. Although real-time reverse transcription polymerase chain reaction (RT-PCR) is known as the most widely laboratory method to detect the COVID-19 from respiratory specimens. It suffers from several main drawbacks such as time-consuming, high false-negative results, and limited availability. Therefore, the automatically detect of COVID-19 will be required.
OBJECTIVE
OBJECTIVE
This study aimed to use an automated deep convolution neural network based pre-trained transfer models for detection of COVID-19 infection in chest X-rays.
MATERIAL AND METHODS
METHODS
In a retrospective study, we have applied Visual Geometry Group (VGG)-16, VGG-19, MobileNet, and InceptionResNetV2 pre-trained models for detection COVID-19 infection from 348 chest X-ray images.
RESULTS
RESULTS
Our proposed models have been trained and tested on a dataset which previously prepared. The all proposed models provide accuracy greater than 90.0%. The pre-trained MobileNet model provides the highest classification performance of automated COVID-19 classification with 99.1% accuracy in comparison with other three proposed models. The plotted area under curve (AUC) of receiver operating characteristics (ROC) of VGG16, VGG19, MobileNet, and InceptionResNetV2 models are 0.92, 0.91, 0.99, and 0.97, respectively.
CONCLUSION
CONCLUSIONS
The all proposed models were able to perform binary classification with the accuracy more than 90.0% for COVID-19 diagnosis. Our data indicated that the MobileNet can be considered as a promising model to detect COVID-19 cases. In the future, by increasing the number of samples of COVID-19 chest X-rays to the training dataset, the accuracy and robustness of our proposed models increase further.
Identifiants
pubmed: 33134214
doi: 10.31661/jbpe.v0i0.2008-1153
pii: JBPE-10-5
pmc: PMC7557468
doi:
Types de publication
Journal Article
Langues
eng
Pagination
559-568Informations de copyright
Copyright: © Journal of Biomedical Physics and Engineering.
Déclaration de conflit d'intérêts
Conflict of Interest: None
Références
Phys Eng Sci Med. 2020 Jun;43(2):635-640
pubmed: 32524445
Phys Eng Sci Med. 2020 Oct 6;:
pubmed: 33025386
Artif Intell Med. 2020 Mar;103:101789
pubmed: 32143796
IET Syst Biol. 2020 Aug;14(4):211-216
pubmed: 32737279
Can J Psychiatry. 2007 Feb;52(2):121-8
pubmed: 17375868
Clin Immunol. 2020 Jun;215:108427
pubmed: 32325252
Radiology. 2020 Aug;296(2):E15-E25
pubmed: 32083985
Radiology. 2020 Aug;296(2):E32-E40
pubmed: 32101510
JAMA. 2020 May 12;323(18):1843-1844
pubmed: 32159775
Radiology. 2020 Aug;296(2):E113-E114
pubmed: 32105562