Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation

Abstract Lung disease analysis in chest X-rays (CXR) using deep learning presents significant challenges due to the wide variation in lung appearance caused by disease progression and differing X-ray settings. While deep learning models have shown remarkable success in segmenting lungs from CXR imag...

Full description

Saved in:
Bibliographic Details
Main Authors: Md. Shariful Alam, Dadong Wang, Yulia Arzhaeva, Jesse Alexander Ende, Joanna Kao, Liz Silverstone, Deborah Yates, Olivier Salvado, Arcot Sowmya
Format: Article
Language:English
Published: Nature Portfolio 2024-11-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-024-79494-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850162524507340800
author Md. Shariful Alam
Dadong Wang
Yulia Arzhaeva
Jesse Alexander Ende
Joanna Kao
Liz Silverstone
Deborah Yates
Olivier Salvado
Arcot Sowmya
author_facet Md. Shariful Alam
Dadong Wang
Yulia Arzhaeva
Jesse Alexander Ende
Joanna Kao
Liz Silverstone
Deborah Yates
Olivier Salvado
Arcot Sowmya
author_sort Md. Shariful Alam
collection DOAJ
description Abstract Lung disease analysis in chest X-rays (CXR) using deep learning presents significant challenges due to the wide variation in lung appearance caused by disease progression and differing X-ray settings. While deep learning models have shown remarkable success in segmenting lungs from CXR images with normal or mildly abnormal findings, their performance declines when faced with complex structures, such as pulmonary opacifications. In this study, we propose AMRU++, an attention-based multi-residual UNet++ network designed for robust and accurate lung segmentation in CXR images with both normal and severe abnormalities. The model incorporates attention modules to capture relevant spatial information and multi-residual blocks to extract rich contextual and discriminative features of lung regions. To further enhance segmentation performance, we introduce a data augmentation technique that simulates the features and characteristics of CXR pathologies, addressing the issue of limited annotated data. Extensive experiments on public and private datasets comprising 350 cases of pneumoconiosis, COVID-19, and tuberculosis validate the effectiveness of our proposed framework and data augmentation technique.
format Article
id doaj-art-35d2b4d8068744e7b639bbf5d220194f
institution OA Journals
issn 2045-2322
language English
publishDate 2024-11-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-35d2b4d8068744e7b639bbf5d220194f2025-08-20T02:22:33ZengNature PortfolioScientific Reports2045-23222024-11-0114111110.1038/s41598-024-79494-wAttention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentationMd. Shariful Alam0Dadong Wang1Yulia Arzhaeva2Jesse Alexander Ende3Joanna Kao4Liz Silverstone5Deborah Yates6Olivier Salvado7Arcot Sowmya8School of Computer Science and Engineering, University of New South WalesCSIRO Data61CSIRO Data61Department of Radiology, St Vincent’s Hospital SydneyDepartment of Radiology, St Vincent’s Hospital SydneyDepartment of Radiology, St Vincent’s Hospital SydneyDepartment of Thoracic Medicine, St Vincent’s Hospital SydneySchool of Electrical Engineering & Robotics, Queensland University of TechnologySchool of Computer Science and Engineering, University of New South WalesAbstract Lung disease analysis in chest X-rays (CXR) using deep learning presents significant challenges due to the wide variation in lung appearance caused by disease progression and differing X-ray settings. While deep learning models have shown remarkable success in segmenting lungs from CXR images with normal or mildly abnormal findings, their performance declines when faced with complex structures, such as pulmonary opacifications. In this study, we propose AMRU++, an attention-based multi-residual UNet++ network designed for robust and accurate lung segmentation in CXR images with both normal and severe abnormalities. The model incorporates attention modules to capture relevant spatial information and multi-residual blocks to extract rich contextual and discriminative features of lung regions. To further enhance segmentation performance, we introduce a data augmentation technique that simulates the features and characteristics of CXR pathologies, addressing the issue of limited annotated data. Extensive experiments on public and private datasets comprising 350 cases of pneumoconiosis, COVID-19, and tuberculosis validate the effectiveness of our proposed framework and data augmentation technique.https://doi.org/10.1038/s41598-024-79494-w
spellingShingle Md. Shariful Alam
Dadong Wang
Yulia Arzhaeva
Jesse Alexander Ende
Joanna Kao
Liz Silverstone
Deborah Yates
Olivier Salvado
Arcot Sowmya
Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation
Scientific Reports
title Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation
title_full Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation
title_fullStr Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation
title_full_unstemmed Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation
title_short Attention-based multi-residual network for lung segmentation in diseased lungs with custom data augmentation
title_sort attention based multi residual network for lung segmentation in diseased lungs with custom data augmentation
url https://doi.org/10.1038/s41598-024-79494-w
work_keys_str_mv AT mdsharifulalam attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT dadongwang attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT yuliaarzhaeva attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT jessealexanderende attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT joannakao attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT lizsilverstone attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT deborahyates attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT oliviersalvado attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation
AT arcotsowmya attentionbasedmultiresidualnetworkforlungsegmentationindiseasedlungswithcustomdataaugmentation