Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs

We present a Machine Learning algorithm based on Python which can be used to aid COVID-19 diagnosis. This algorithm employs Convolutional Neural Networks (CNN) of ResNet-18 architecture from thoracic X-ray images to build a trained dataset that enables further comparisons between common pulmonary di...

Full description

Saved in:
Bibliographic Details
Main Authors: Bruno Takara, Felipe Freitas, Alexandre Bacelar, Rochelle Lykawka, Mirko Salomon Alva Sanchez
Format: Article
Language:English
Published: Brazilian Radiation Protection Society (Sociedade Brasileira de Proteção Radiológica, SBPR) 2022-09-01
Series:Brazilian Journal of Radiation Sciences
Subjects:
Online Access:https://bjrs.org.br/revista/index.php/REVISTA/article/view/2056
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849318474864656384
author Bruno Takara
Felipe Freitas
Alexandre Bacelar
Rochelle Lykawka
Mirko Salomon Alva Sanchez
author_facet Bruno Takara
Felipe Freitas
Alexandre Bacelar
Rochelle Lykawka
Mirko Salomon Alva Sanchez
author_sort Bruno Takara
collection DOAJ
description We present a Machine Learning algorithm based on Python which can be used to aid COVID-19 diagnosis. This algorithm employs Convolutional Neural Networks (CNN) of ResNet-18 architecture from thoracic X-ray images to build a trained dataset that enables further comparisons between common pulmonary diseases and COVID-19 diagnosed patients to classify the radiological findings as being due the COVID-19 or other pathologies. We discuss the importance of setting the right parameters related to training and what they might represent in clinical procedures. We used a dataset containing 942 COVID-19 labeled radiographs from HCPA - Hospital das Clínicas de Porto Alegre and compared it to a public dataset from NIH Clinical Center containing images of pulmonary diseases. Lastly, our trained model had an accuracy of 81.76% for the imbalanced classes and an accuracy of 46.94% for the balanced classes, when compared to other pulmonary diseases such as pneumonia, edema, mass, consolidation, and fibrosis. These results disclose the difficulty of diagnosing COVID-19 from a chest radiograph as it resembles other pulmonary illnesses and makes room for further research in this matter.
format Article
id doaj-art-3a07564049364773a42ffc73e8f2cc46
institution Kabale University
issn 2319-0612
language English
publishDate 2022-09-01
publisher Brazilian Radiation Protection Society (Sociedade Brasileira de Proteção Radiológica, SBPR)
record_format Article
series Brazilian Journal of Radiation Sciences
spelling doaj-art-3a07564049364773a42ffc73e8f2cc462025-08-20T03:50:49ZengBrazilian Radiation Protection Society (Sociedade Brasileira de Proteção Radiológica, SBPR)Brazilian Journal of Radiation Sciences2319-06122022-09-0110310.15392/bjrs.v10i3.20561674Artificial intelligence to evaluate diagnosed COVID-19 chest radiographsBruno Takara0Felipe FreitasAlexandre BacelarRochelle LykawkaMirko Salomon Alva Sanchez1Universidade Federal de Ciências da Saúde de Porto Alegre, Porto Alegre, Rio Grande do Sul, BrazilUFCSPAWe present a Machine Learning algorithm based on Python which can be used to aid COVID-19 diagnosis. This algorithm employs Convolutional Neural Networks (CNN) of ResNet-18 architecture from thoracic X-ray images to build a trained dataset that enables further comparisons between common pulmonary diseases and COVID-19 diagnosed patients to classify the radiological findings as being due the COVID-19 or other pathologies. We discuss the importance of setting the right parameters related to training and what they might represent in clinical procedures. We used a dataset containing 942 COVID-19 labeled radiographs from HCPA - Hospital das Clínicas de Porto Alegre and compared it to a public dataset from NIH Clinical Center containing images of pulmonary diseases. Lastly, our trained model had an accuracy of 81.76% for the imbalanced classes and an accuracy of 46.94% for the balanced classes, when compared to other pulmonary diseases such as pneumonia, edema, mass, consolidation, and fibrosis. These results disclose the difficulty of diagnosing COVID-19 from a chest radiograph as it resembles other pulmonary illnesses and makes room for further research in this matter.https://bjrs.org.br/revista/index.php/REVISTA/article/view/2056x-rayartificial inteligenceradiography
spellingShingle Bruno Takara
Felipe Freitas
Alexandre Bacelar
Rochelle Lykawka
Mirko Salomon Alva Sanchez
Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs
Brazilian Journal of Radiation Sciences
x-ray
artificial inteligence
radiography
title Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs
title_full Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs
title_fullStr Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs
title_full_unstemmed Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs
title_short Artificial intelligence to evaluate diagnosed COVID-19 chest radiographs
title_sort artificial intelligence to evaluate diagnosed covid 19 chest radiographs
topic x-ray
artificial inteligence
radiography
url https://bjrs.org.br/revista/index.php/REVISTA/article/view/2056
work_keys_str_mv AT brunotakara artificialintelligencetoevaluatediagnosedcovid19chestradiographs
AT felipefreitas artificialintelligencetoevaluatediagnosedcovid19chestradiographs
AT alexandrebacelar artificialintelligencetoevaluatediagnosedcovid19chestradiographs
AT rochellelykawka artificialintelligencetoevaluatediagnosedcovid19chestradiographs
AT mirkosalomonalvasanchez artificialintelligencetoevaluatediagnosedcovid19chestradiographs