TY - JOUR
T1 - DML-OFA
T2 - Deep mutual learning with online feature alignment for the detection of COVID-19 from chest x-ray images
AU - Liang, Zhihao
AU - Lu, Huijuan
AU - Ming, Zhendong
AU - Chai, Zhuijun
AU - Yao, Yudong
N1 - Publisher Copyright:
© 2024 John Wiley & Sons Ltd.
PY - 2024/5/15
Y1 - 2024/5/15
N2 - 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 (Formula presented.) Accuracy, 97.28 (Formula presented.) Specificity, 96.21 (Formula presented.) Recall, 97.45 (Formula presented.) Precision, and 96.82 (Formula presented.) F1-score, which outperforms other previous related works.
AB - 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 (Formula presented.) Accuracy, 97.28 (Formula presented.) Specificity, 96.21 (Formula presented.) Recall, 97.45 (Formula presented.) Precision, and 96.82 (Formula presented.) F1-score, which outperforms other previous related works.
KW - COVID-19 recognition
KW - deep mutual learning
KW - feature fusion
UR - http://www.scopus.com/inward/record.url?scp=85182841048&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85182841048&partnerID=8YFLogxK
U2 - 10.1002/cpe.8023
DO - 10.1002/cpe.8023
M3 - Article
AN - SCOPUS:85182841048
SN - 1532-0626
VL - 36
JO - Concurrency and Computation: Practice and Experience
JF - Concurrency and Computation: Practice and Experience
IS - 11
M1 - e8023
ER -