DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images
Summary COVID‐19 is a novel coronavirus‐induced disease and automatic identification of COVID‐19 using computer‐assisted methods can facilitate faster diagnostic efficiency. Current research typically employs a single model for COVID‐19 identification, while implicit and complementary knowledge betw...
Published in: | Concurrency and Computation: Practice and Experience |
---|---|
Main Authors: | , , , , |
Other Authors: | , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Wiley
2024
|
Subjects: | |
Online Access: | http://dx.doi.org/10.1002/cpe.8023 https://onlinelibrary.wiley.com/doi/pdf/10.1002/cpe.8023 |
id |
crwiley:10.1002/cpe.8023 |
---|---|
record_format |
openpolar |
spelling |
crwiley:10.1002/cpe.8023 2024-06-02T08:05:47+00:00 DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images Liang, Zhihao Lu, Huijuan Ming, Zhendong Chai, Zhuijun Yao, Yudong Science and Technology Program of Zhejiang Province Natural Science Foundation of Zhejiang Province National Natural Science Foundation of China 2024 http://dx.doi.org/10.1002/cpe.8023 https://onlinelibrary.wiley.com/doi/pdf/10.1002/cpe.8023 en eng Wiley http://onlinelibrary.wiley.com/termsAndConditions#vor Concurrency and Computation: Practice and Experience volume 36, issue 11 ISSN 1532-0626 1532-0634 journal-article 2024 crwiley https://doi.org/10.1002/cpe.8023 2024-05-03T11:04:53Z Summary COVID‐19 is a novel coronavirus‐induced disease and automatic identification of COVID‐19 using computer‐assisted methods can facilitate faster diagnostic efficiency. Current research typically employs a single model for COVID‐19 identification, while implicit and complementary knowledge between heterogeneous networks is neglected. To address these issues, we propose a new model based on deep mutual learning with online feature alignment called DML‐OFA to more effectively diagnose COVID‐19. First, we use a traditional deep mutual learning (DML) framework to allow two parallel heterogeneous networks to learn from each other to form two effective feature extractors. In addition, we embed the adaptive feature fusion classifier and logits ensembling module in the proposed DML‐OFA, which can simultaneously learn implicit complementary knowledge from feature maps and logits. We evaluated DML‐OFA on four public datasets: Covid‐chestxray‐dataset, ChestXRay2017, Coronavirus‐dataset and COVIDx. The results showed that our model attains 97.10 Accuracy, 97.28 Specificity, 96.21 Recall, 97.45 Precision, and 96.82 F1‐score, which outperforms other previous related works. Article in Journal/Newspaper DML Wiley Online Library Concurrency and Computation: Practice and Experience |
institution |
Open Polar |
collection |
Wiley Online Library |
op_collection_id |
crwiley |
language |
English |
description |
Summary COVID‐19 is a novel coronavirus‐induced disease and automatic identification of COVID‐19 using computer‐assisted methods can facilitate faster diagnostic efficiency. Current research typically employs a single model for COVID‐19 identification, while implicit and complementary knowledge between heterogeneous networks is neglected. To address these issues, we propose a new model based on deep mutual learning with online feature alignment called DML‐OFA to more effectively diagnose COVID‐19. First, we use a traditional deep mutual learning (DML) framework to allow two parallel heterogeneous networks to learn from each other to form two effective feature extractors. In addition, we embed the adaptive feature fusion classifier and logits ensembling module in the proposed DML‐OFA, which can simultaneously learn implicit complementary knowledge from feature maps and logits. We evaluated DML‐OFA on four public datasets: Covid‐chestxray‐dataset, ChestXRay2017, Coronavirus‐dataset and COVIDx. The results showed that our model attains 97.10 Accuracy, 97.28 Specificity, 96.21 Recall, 97.45 Precision, and 96.82 F1‐score, which outperforms other previous related works. |
author2 |
Science and Technology Program of Zhejiang Province Natural Science Foundation of Zhejiang Province National Natural Science Foundation of China |
format |
Article in Journal/Newspaper |
author |
Liang, Zhihao Lu, Huijuan Ming, Zhendong Chai, Zhuijun Yao, Yudong |
spellingShingle |
Liang, Zhihao Lu, Huijuan Ming, Zhendong Chai, Zhuijun Yao, Yudong DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images |
author_facet |
Liang, Zhihao Lu, Huijuan Ming, Zhendong Chai, Zhuijun Yao, Yudong |
author_sort |
Liang, Zhihao |
title |
DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images |
title_short |
DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images |
title_full |
DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images |
title_fullStr |
DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images |
title_full_unstemmed |
DML‐OFA: Deep mutual learning with online feature alignment for the detection of COVID‐19 from chest x‐ray images |
title_sort |
dml‐ofa: deep mutual learning with online feature alignment for the detection of covid‐19 from chest x‐ray images |
publisher |
Wiley |
publishDate |
2024 |
url |
http://dx.doi.org/10.1002/cpe.8023 https://onlinelibrary.wiley.com/doi/pdf/10.1002/cpe.8023 |
genre |
DML |
genre_facet |
DML |
op_source |
Concurrency and Computation: Practice and Experience volume 36, issue 11 ISSN 1532-0626 1532-0634 |
op_rights |
http://onlinelibrary.wiley.com/termsAndConditions#vor |
op_doi |
https://doi.org/10.1002/cpe.8023 |
container_title |
Concurrency and Computation: Practice and Experience |
_version_ |
1800750659830546432 |