Cover Image

Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.

COVID-19 has been one of the popular foci in the research community since its first outbreak in China, 2019. Radiological patterns such as ground glass opacity (GGO) and consolidations are often found in CT scan images of moderate to severe COVID-19 patients. Therefore, a deep learning model can be...

Full description

Saved in:
Bibliographic Details
Main Authors: Thon, Pun Liang, Than, Joel C. M., M. Noor, Norliza, Han, Jun, Then, Patrick
Format: Article
Language:English
Published: Penerbit UTHM 2023
Subjects:
TK Electrical engineering. Electronics Nuclear engineering
Online Access:http://eprints.utm.my/105723/1/NorlizaMNoor2023_InvestigationofConVitonCovid19LungImage.pdf
http://eprints.utm.my/105723/
http://dx.doi.org/10.30880/ijie.2023.15.03.005
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.utm.105723
record_format eprints
spelling my.utm.1057232024-05-13T07:20:14Z http://eprints.utm.my/105723/ Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads. Thon, Pun Liang Than, Joel C. M. M. Noor, Norliza Han, Jun Then, Patrick TK Electrical engineering. Electronics Nuclear engineering COVID-19 has been one of the popular foci in the research community since its first outbreak in China, 2019. Radiological patterns such as ground glass opacity (GGO) and consolidations are often found in CT scan images of moderate to severe COVID-19 patients. Therefore, a deep learning model can be trained to distinguish COVID-19 patients using their CT scan images. Convolutional Neural Networks (CNNs) has been a popular choice for this type of classification task. Another potential method is the use of vision transformer with convolution, resulting in Convolutional Vision Transformer (ConViT), to possibly produce on par performance using less computational resources. In this study, ConViT is applied to diagnose COVID-19 cases from lung CT scan images. Particularly, we investigated the relationship of the input image pixel resolutions and the number of attention heads used in ConViT and their effects on the model’s performance. Specifically, we used 512x512, 224x224 and 128x128 pixels resolution to train the model with 4 (tiny), 9 (small) and 16 (base) number of attention heads used. An open access dataset consisting of 2282 COVID-19 CT images and 9776 Normal CT images from Iran is used in this study. By using 128x128 image pixels resolution, training using 16 attention heads, the ConViT model has achieved an accuracy of 98.01%, sensitivity of 90.83%, specificity of 99.69%, positive predictive value (PPV) of 95.58%, negative predictive value (NPV) of 97.89% and F1-score of 94.55%. The model has also achieved improved performance over other recent studies that used the same dataset. In conclusion, this study has shown that the ConViT model can play a meaningful role to complement RT-PCR test on COVID-19 close contacts and patients. Penerbit UTHM 2023-07-31 Article PeerReviewed application/pdf en http://eprints.utm.my/105723/1/NorlizaMNoor2023_InvestigationofConVitonCovid19LungImage.pdf Thon, Pun Liang and Than, Joel C. M. and M. Noor, Norliza and Han, Jun and Then, Patrick (2023) Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads. International Journal of Integrated Engineering, 15 (3). pp. 54-63. ISSN 2229-838X http://dx.doi.org/10.30880/ijie.2023.15.03.005 DOI: 10.30880/ijie.2023.15.03.005
institution Universiti Teknologi Malaysia
building UTM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Teknologi Malaysia
content_source UTM Institutional Repository
url_provider http://eprints.utm.my/
language English
topic TK Electrical engineering. Electronics Nuclear engineering
spellingShingle TK Electrical engineering. Electronics Nuclear engineering
Thon, Pun Liang
Than, Joel C. M.
M. Noor, Norliza
Han, Jun
Then, Patrick
Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.
description COVID-19 has been one of the popular foci in the research community since its first outbreak in China, 2019. Radiological patterns such as ground glass opacity (GGO) and consolidations are often found in CT scan images of moderate to severe COVID-19 patients. Therefore, a deep learning model can be trained to distinguish COVID-19 patients using their CT scan images. Convolutional Neural Networks (CNNs) has been a popular choice for this type of classification task. Another potential method is the use of vision transformer with convolution, resulting in Convolutional Vision Transformer (ConViT), to possibly produce on par performance using less computational resources. In this study, ConViT is applied to diagnose COVID-19 cases from lung CT scan images. Particularly, we investigated the relationship of the input image pixel resolutions and the number of attention heads used in ConViT and their effects on the model’s performance. Specifically, we used 512x512, 224x224 and 128x128 pixels resolution to train the model with 4 (tiny), 9 (small) and 16 (base) number of attention heads used. An open access dataset consisting of 2282 COVID-19 CT images and 9776 Normal CT images from Iran is used in this study. By using 128x128 image pixels resolution, training using 16 attention heads, the ConViT model has achieved an accuracy of 98.01%, sensitivity of 90.83%, specificity of 99.69%, positive predictive value (PPV) of 95.58%, negative predictive value (NPV) of 97.89% and F1-score of 94.55%. The model has also achieved improved performance over other recent studies that used the same dataset. In conclusion, this study has shown that the ConViT model can play a meaningful role to complement RT-PCR test on COVID-19 close contacts and patients.
format Article
author Thon, Pun Liang
Than, Joel C. M.
M. Noor, Norliza
Han, Jun
Then, Patrick
author_facet Thon, Pun Liang
Than, Joel C. M.
M. Noor, Norliza
Han, Jun
Then, Patrick
author_sort Thon, Pun Liang
title Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.
title_short Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.
title_full Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.
title_fullStr Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.
title_full_unstemmed Investigation of ConViT on COVID-19 lung image classification and the effects of image resolution and number of attention heads.
title_sort investigation of convit on covid-19 lung image classification and the effects of image resolution and number of attention heads.
publisher Penerbit UTHM
publishDate 2023
url http://eprints.utm.my/105723/1/NorlizaMNoor2023_InvestigationofConVitonCovid19LungImage.pdf
http://eprints.utm.my/105723/
http://dx.doi.org/10.30880/ijie.2023.15.03.005
_version_ 1800082654056415232
score 13.211869

Similar Items