COVID-19 Detection via Image Classification using Deep Learning on Chest X-Ray

Mohammad Ayyaz Azeem; Muhammad Irfan Khan; Suleiman A. Khan

doi:10.1109/EE-RDS53766.2021.9708588

COVID-19 Detection via Image Classification using Deep Learning on Chest X-Ray

Mohammad Ayyaz Azeem, Muhammad Irfan Khan, Suleiman A. Khan

Suomen molekyylilääketieteen instituutti

Tutkimustuotos: Artikkeli kirjassa/raportissa/konferenssijulkaisussa › Konferenssiartikkeli › Tieteellinen › vertaisarvioitu

Abstrakti

The focus of this study is to evaluate and examine a set of deep learning transfer learning techniques applied to chest radiograph images for the classification of COVID-19, normal (healthy), and pneumonia. In this work, we have used four transfer learning models, VGG16, InceptionV3, ResNet50, and DenseNet121 for the classification tasks. Our results indicate that the VGG16 method outperforms comparative classification models in terms of accuracy, sensitivity, and specificity. The VGG16 model detects and classifies COVID-19, normal (healthy), and pneumonia with 94% test accuracy, 94% sensitivity, and 94.20% specificity. Code is publically available at: https://github.com/ayyaz-azeem/Covid19challenge.git

Alkuperäiskieli	englanti
Otsikko	2021 Ethics and Explainability for Responsible Data Science (EE-RDS)
Kustantaja	IEEE
Julkaisupäivä	2021
ISBN (elektroninen)	978-1-6654-8358-2
DOI - pysyväislinkit	https://doi.org/10.1109/EE-RDS53766.2021.9708588
Tila	Julkaistu - 2021
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisuussa
Tapahtuma	Ethics and Explainability for Responsible Data Science Conference - Johannesburg, Etelä-Afrikka Kesto: 27 lokak. 2021 → 28 lokak. 2021

Lisätietoja

Publisher Copyright:
© 2021 IEEE.

Tieteenalat

113 Tietojenkäsittely- ja informaatiotieteet

Pääsy asiakirjaan

10.1109/EE-RDS53766.2021.9708588

Siteeraa tätä

@inproceedings{d7fd60c86f864c05bddd0ac87886ab0d,

title = "COVID-19 Detection via Image Classification using Deep Learning on Chest X-Ray",

abstract = "The focus of this study is to evaluate and examine a set of deep learning transfer learning techniques applied to chest radiograph images for the classification of COVID-19, normal (healthy), and pneumonia. In this work, we have used four transfer learning models, VGG16, InceptionV3, ResNet50, and DenseNet121 for the classification tasks. Our results indicate that the VGG16 method outperforms comparative classification models in terms of accuracy, sensitivity, and specificity. The VGG16 model detects and classifies COVID-19, normal (healthy), and pneumonia with 94% test accuracy, 94% sensitivity, and 94.20% specificity. Code is publically available at: https://github.com/ayyaz-azeem/Covid19challenge.git",

keywords = "COVID-19 classification, COVID-19 detection, CXR images, Deep Learning, Machine Learning, Transfer Learning, VGG16, X-Ray, 113 Computer and information sciences",

author = "Azeem, {Mohammad Ayyaz} and Khan, {Muhammad Irfan} and Khan, {Suleiman A.}",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; Ethics and Explainability for Responsible Data Science Conference, EE-RDS ; Conference date: 27-10-2021 Through 28-10-2021",

year = "2021",

doi = "10.1109/EE-RDS53766.2021.9708588",

language = "English",

booktitle = "2021 Ethics and Explainability for Responsible Data Science (EE-RDS)",

publisher = "IEEE",

address = "United States",

}

TY - GEN

T1 - COVID-19 Detection via Image Classification using Deep Learning on Chest X-Ray

AU - Azeem, Mohammad Ayyaz

AU - Khan, Muhammad Irfan

AU - Khan, Suleiman A.

PY - 2021

Y1 - 2021

N2 - The focus of this study is to evaluate and examine a set of deep learning transfer learning techniques applied to chest radiograph images for the classification of COVID-19, normal (healthy), and pneumonia. In this work, we have used four transfer learning models, VGG16, InceptionV3, ResNet50, and DenseNet121 for the classification tasks. Our results indicate that the VGG16 method outperforms comparative classification models in terms of accuracy, sensitivity, and specificity. The VGG16 model detects and classifies COVID-19, normal (healthy), and pneumonia with 94% test accuracy, 94% sensitivity, and 94.20% specificity. Code is publically available at: https://github.com/ayyaz-azeem/Covid19challenge.git

AB - The focus of this study is to evaluate and examine a set of deep learning transfer learning techniques applied to chest radiograph images for the classification of COVID-19, normal (healthy), and pneumonia. In this work, we have used four transfer learning models, VGG16, InceptionV3, ResNet50, and DenseNet121 for the classification tasks. Our results indicate that the VGG16 method outperforms comparative classification models in terms of accuracy, sensitivity, and specificity. The VGG16 model detects and classifies COVID-19, normal (healthy), and pneumonia with 94% test accuracy, 94% sensitivity, and 94.20% specificity. Code is publically available at: https://github.com/ayyaz-azeem/Covid19challenge.git

KW - COVID-19 classification

KW - COVID-19 detection

KW - CXR images

KW - Deep Learning

KW - Machine Learning

KW - Transfer Learning

KW - VGG16

KW - X-Ray

KW - 113 Computer and information sciences

U2 - 10.1109/EE-RDS53766.2021.9708588

DO - 10.1109/EE-RDS53766.2021.9708588

M3 - Conference contribution

AN - SCOPUS:85125933056

BT - 2021 Ethics and Explainability for Responsible Data Science (EE-RDS)

PB - IEEE

T2 - Ethics and Explainability for Responsible Data Science Conference

Y2 - 27 October 2021 through 28 October 2021

ER -