Download
journal.pone.0247839.pdf 2,94MB
WeightNameValue
1000 Titel
  • Convolutional neural network model based on radiological images to support COVID-19 diagnosis: Evaluating database biases
1000 Autor/in
  1. Maior, Caio B. S. |
  2. Santana, João Mateus |
  3. Lins, Isis D. |
  4. Moura, Márcio J. C. |
1000 Erscheinungsjahr 2021
1000 Publikationstyp
  1. Artikel |
1000 Online veröffentlicht
  • 2021-03-01
1000 Erschienen in
1000 Quellenangabe
  • 16(3):e0247839
1000 Copyrightjahr
  • 2021
1000 Lizenz
1000 Verlagsversion
  • https://doi.org/10.1371/journal.pone.0247839 |
1000 Ergänzendes Material
  • https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0247839#sec026 |
1000 Publikationsstatus
1000 Begutachtungsstatus
1000 Sprache der Publikation
1000 Abstract/Summary
  • As SARS-CoV-2 has spread quickly throughout the world, the scientific community has spent major efforts on better understanding the characteristics of the virus and possible means to prevent, diagnose, and treat COVID-19. A valid approach presented in the literature is to develop an image-based method to support COVID-19 diagnosis using convolutional neural networks (CNN). Because the availability of radiological data is rather limited due to the novelty of COVID-19, several methodologies consider reduced datasets, which may be inadequate, biasing the model. Here, we performed an analysis combining six different databases using chest X-ray images from open datasets to distinguish images of infected patients while differentiating COVID-19 and pneumonia from ‘no-findings’ images. In addition, the performance of models created from fewer databases, which may imperceptibly overestimate their results, is discussed. Two CNN-based architectures were created to process images of different sizes (512 × 512, 768 × 768, 1024 × 1024, and 1536 × 1536). Our best model achieved a balanced accuracy (BA) of 87.7% in predicting one of the three classes (‘no-findings’, ‘COVID-19’, and ‘pneumonia’) and a specific balanced precision of 97.0% for ‘COVID-19’ class. We also provided binary classification with a precision of 91.0% for detection of sick patients (i.e., with COVID-19 or pneumonia) and 98.4% for COVID-19 detection (i.e., differentiating from ‘no-findings’ or ‘pneumonia’). Indeed, despite we achieved an unrealistic 97.2% BA performance for one specific case, the proposed methodology of using multiple databases achieved better and less inflated results than from models with specific image datasets for training. Thus, this framework is promising for a low-cost, fast, and noninvasive means to support the diagnosis of COVID-19.
1000 Sacherschließung
gnd 1206347392 COVID-19
lokal Virus testing
lokal Database and informatics methods
lokal Computed axial tomography
lokal Pneumonia
lokal Imaging techniques
lokal Radiology and imaging
lokal X-ray radiography
1000 Fächerklassifikation (DDC)
1000 Liste der Beteiligten
  1. https://frl.publisso.de/adhoc/uri/TWFpb3IsIENhaW8gQi4gUy4=|https://orcid.org/0000-0001-7666-4928|https://frl.publisso.de/adhoc/uri/TGlucywgSXNpcyBELg==|https://frl.publisso.de/adhoc/uri/TW91cmEsIE3DoXJjaW8gSi4gQy4=
1000 Label
1000 Förderer
  1. Universidade Federal de Pernambuco |
  2. Conselho Nacional de Desenvolvimento Científico e Tecnológico |
  3. Coordenação de Aperfeiçoamento de Pessoal de Nível Superior |
1000 Fördernummer
  1. 23076.019632/2020-11
  2. 305696/2018-1; 309617/2019-7
  3. 001
1000 Förderprogramm
  1. Pró-Reitoria de Pesquisa e Inovação (Propesqi)
  2. -
  3. -
1000 Dateien
1000 Förderung
  1. 1000 joinedFunding-child
    1000 Förderer Universidade Federal de Pernambuco |
    1000 Förderprogramm Pró-Reitoria de Pesquisa e Inovação (Propesqi)
    1000 Fördernummer 23076.019632/2020-11
  2. 1000 joinedFunding-child
    1000 Förderer Conselho Nacional de Desenvolvimento Científico e Tecnológico |
    1000 Förderprogramm -
    1000 Fördernummer 305696/2018-1; 309617/2019-7
  3. 1000 joinedFunding-child
    1000 Förderer Coordenação de Aperfeiçoamento de Pessoal de Nível Superior |
    1000 Förderprogramm -
    1000 Fördernummer 001
1000 Objektart article
1000 Beschrieben durch
1000 @id frl:6425875.rdf
1000 Erstellt am 2021-03-02T11:29:53.176+0100
1000 Erstellt von 122
1000 beschreibt frl:6425875
1000 Bearbeitet von 122
1000 Zuletzt bearbeitet Tue Mar 02 11:32:02 CET 2021
1000 Objekt bearb. Tue Mar 02 11:31:35 CET 2021
1000 Vgl. frl:6425875
1000 Oai Id
  1. oai:frl.publisso.de:frl:6425875 |
  2. oai:frl.publisso.de:frl:6425875 |
1000 Sichtbarkeit Metadaten public
1000 Sichtbarkeit Daten public
1000 Gegenstand von

View source