File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Deep learning attention-guided radiomics for COVID-19 chest radiograph classification

TitleDeep learning attention-guided radiomics for COVID-19 chest radiograph classification
Authors
Keywordschest radiograph
classification
Coronavirus disease 2019 (COVID-19)
deep learning
radiomics
Issue Date1-Feb-2023
PublisherAME Publishing
Citation
Quantitative Imaging in Medicine and Surgery, 2023, v. 13, n. 2, p. 572-584 How to Cite?
Abstract

Background: Accurate assessment of coronavirus disease 2019 (COVID-19) lung involvement through chest radiograph plays an important role in effective management of the infection. This study aims to develop a two-step feature merging method to integrate image features from deep learning and radiomics to differentiate COVID-19, non-COVID-19 pneumonia and normal chest radiographs (CXR).

Methods: In this study, a deformable convolutional neural network (deformable CNN) was developed and used as a feature extractor to obtain 1,024-dimensional deep learning latent representation (DLR) features. Then 1,069-dimensional radiomics features were extracted from the region of interest (ROI) guided by deformable CNN’s attention. The two feature sets were concatenated to generate a merged feature set for classification. For comparative experiments, the same process has been applied to the DLR-only feature set for verifying the effectiveness of feature concatenation.

Results: Using the merged feature set resulted in an overall average accuracy of 91.0% for three-class classification, representing a statistically significant improvement of 0.6% compared to the DLR-only classification. The recall and precision of classification into the COVID-19 class were 0.926 and 0.976, respectively. The feature merging method was shown to significantly improve the classification performance as compared to using only deep learning features, regardless of choice of classifier (P value <0.0001). Three classes’ F1-score were 0.892, 0.890, and 0.950 correspondingly (i.e., normal, non-COVID-19 pneumonia, COVID-19).

Conclusions: A two-step COVID-19 classification framework integrating information from both DLR and radiomics features (guided by deep learning attention mechanism) has been developed. The proposed feature merging method has been shown to improve the performance of chest radiograph classification as compared to the case of using only deep learning features.


Persistent Identifierhttp://hdl.handle.net/10722/338905
ISSN
2021 Impact Factor: 4.630
2020 SCImago Journal Rankings: 0.766
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorYang, DR-
dc.contributor.authorRen, G-
dc.contributor.authorNi, RY-
dc.contributor.authorHuang, YH-
dc.contributor.authorLam, NFD-
dc.contributor.authorSun, HF-
dc.contributor.authorWan, SBN-
dc.contributor.authorWong, MFE-
dc.contributor.authorChan, KK-
dc.contributor.authorTsang, HCH-
dc.contributor.authorXu, L-
dc.contributor.authorWu, TC-
dc.contributor.authorKong, FM-
dc.contributor.authorWáng, YXJ-
dc.contributor.authorQin, J-
dc.contributor.authorChan, LWC-
dc.contributor.authorYing, M-
dc.contributor.authorCai, J-
dc.date.accessioned2024-03-11T10:32:25Z-
dc.date.available2024-03-11T10:32:25Z-
dc.date.issued2023-02-01-
dc.identifier.citationQuantitative Imaging in Medicine and Surgery, 2023, v. 13, n. 2, p. 572-584-
dc.identifier.issn2223-4292-
dc.identifier.urihttp://hdl.handle.net/10722/338905-
dc.description.abstract<p><strong>Background: </strong>Accurate assessment of coronavirus disease 2019 (COVID-19) lung involvement through chest radiograph plays an important role in effective management of the infection. This study aims to develop a two-step feature merging method to integrate image features from deep learning and radiomics to differentiate COVID-19, non-COVID-19 pneumonia and normal chest radiographs (CXR).</p><p><strong>Methods: </strong>In this study, a deformable convolutional neural network (deformable CNN) was developed and used as a feature extractor to obtain 1,024-dimensional deep learning latent representation (DLR) features. Then 1,069-dimensional radiomics features were extracted from the region of interest (ROI) guided by deformable CNN’s attention. The two feature sets were concatenated to generate a merged feature set for classification. For comparative experiments, the same process has been applied to the DLR-only feature set for verifying the effectiveness of feature concatenation.</p><p><strong>Results: </strong>Using the merged feature set resulted in an overall average accuracy of 91.0% for three-class classification, representing a statistically significant improvement of 0.6% compared to the DLR-only classification. The recall and precision of classification into the COVID-19 class were 0.926 and 0.976, respectively. The feature merging method was shown to significantly improve the classification performance as compared to using only deep learning features, regardless of choice of classifier (P value <0.0001). Three classes’ F1-score were 0.892, 0.890, and 0.950 correspondingly (i.e., normal, non-COVID-19 pneumonia, COVID-19).</p><p><strong>Conclusions: </strong>A two-step COVID-19 classification framework integrating information from both DLR and radiomics features (guided by deep learning attention mechanism) has been developed. The proposed feature merging method has been shown to improve the performance of chest radiograph classification as compared to the case of using only deep learning features.</p>-
dc.languageeng-
dc.publisherAME Publishing-
dc.relation.ispartofQuantitative Imaging in Medicine and Surgery-
dc.rightsThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.-
dc.subjectchest radiograph-
dc.subjectclassification-
dc.subjectCoronavirus disease 2019 (COVID-19)-
dc.subjectdeep learning-
dc.subjectradiomics-
dc.titleDeep learning attention-guided radiomics for COVID-19 chest radiograph classification-
dc.typeArticle-
dc.identifier.doi10.21037/qims-22-531-
dc.identifier.scopuseid_2-s2.0-85147155108-
dc.identifier.volume13-
dc.identifier.issue2-
dc.identifier.spage572-
dc.identifier.epage584-
dc.identifier.eissn2223-4306-
dc.identifier.isiWOS:000890272900001-
dc.identifier.issnl2223-4306-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats