DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT)
Tuberculosis (TB) remains a significant global health challenge, characterized by high incidence and mortality rates on a global scale. With the rapid advancement of computer-aided diagnosis (CAD) tools in recent years, CAD has assumed an increasingly crucial role in supporting TB diagnosis. Nonethe...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2024-02-01
|
Series: | Heliyon |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S2405844024015214 |
_version_ | 1797267851366629376 |
---|---|
author | Zhuoyi Tan Hizmawati Madzin Bahari Norafida Yang ChongShuang Wei Sun Tianyu Nie Fengzhou Cai |
author_facet | Zhuoyi Tan Hizmawati Madzin Bahari Norafida Yang ChongShuang Wei Sun Tianyu Nie Fengzhou Cai |
author_sort | Zhuoyi Tan |
collection | DOAJ |
description | Tuberculosis (TB) remains a significant global health challenge, characterized by high incidence and mortality rates on a global scale. With the rapid advancement of computer-aided diagnosis (CAD) tools in recent years, CAD has assumed an increasingly crucial role in supporting TB diagnosis. Nonetheless, the development of CAD for TB diagnosis heavily relies on well-annotated computerized tomography (CT) datasets. Currently, the available annotations in TB CT datasets are still limited, which in turn restricts the development of CAD tools for TB diagnosis to some extent. To address this limitation, we introduce DeepPulmoTB, a CT multi-task learning dataset explicitly designed for TB diagnosis. To demonstrate the advantages of DeepPulmoTB, we propose a novel multi-task learning model, DeepPulmoTBNet (DPTBNet), for the joint segmentation and classification of lesion tissues in CT images. The architecture of DPTBNet comprises two subnets: SwinUnetR for the segmentation task, and a lightweight multi-scale network for the classification task. Furthermore, to enhance the model's capacity to capture TB lesion features, we introduce an improved iterative optimization algorithm that refines feature maps by integrating probability maps obtained in previous iterations. Extensive experiments validate the effectiveness of DPTBNet and the practicality of the DeepPulmoTB dataset. |
first_indexed | 2024-03-08T03:30:49Z |
format | Article |
id | doaj.art-eee36a8c8e79410d860a0ae570e481a1 |
institution | Directory Open Access Journal |
issn | 2405-8440 |
language | English |
last_indexed | 2024-04-25T01:23:09Z |
publishDate | 2024-02-01 |
publisher | Elsevier |
record_format | Article |
series | Heliyon |
spelling | doaj.art-eee36a8c8e79410d860a0ae570e481a12024-03-09T09:25:27ZengElsevierHeliyon2405-84402024-02-01104e25490DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT)Zhuoyi Tan0Hizmawati Madzin1Bahari Norafida2Yang ChongShuang3Wei Sun4Tianyu Nie5Fengzhou Cai6Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, Serdang, 43400, MalaysiaFaculty of Computer Science and Information Technology, Universiti Putra Malaysia, Serdang, 43400, Malaysia; Corresponding author.Department of Radiology, Universit Putra Malaysia, 43400 Serdang, Selangor, MalaysiaDepartment of Radiology, Universit Putra Malaysia, 43400 Serdang, Selangor, MalaysiaFaculty of Computer Science and Information Technology, Universiti Putra Malaysia, Serdang, 43400, MalaysiaCollege of Computer Science, Chongqing University, Chongqing 400030, ChinaElectronics and Computer Science, University of Southampton, Southampton, SO17 1BJ, UKTuberculosis (TB) remains a significant global health challenge, characterized by high incidence and mortality rates on a global scale. With the rapid advancement of computer-aided diagnosis (CAD) tools in recent years, CAD has assumed an increasingly crucial role in supporting TB diagnosis. Nonetheless, the development of CAD for TB diagnosis heavily relies on well-annotated computerized tomography (CT) datasets. Currently, the available annotations in TB CT datasets are still limited, which in turn restricts the development of CAD tools for TB diagnosis to some extent. To address this limitation, we introduce DeepPulmoTB, a CT multi-task learning dataset explicitly designed for TB diagnosis. To demonstrate the advantages of DeepPulmoTB, we propose a novel multi-task learning model, DeepPulmoTBNet (DPTBNet), for the joint segmentation and classification of lesion tissues in CT images. The architecture of DPTBNet comprises two subnets: SwinUnetR for the segmentation task, and a lightweight multi-scale network for the classification task. Furthermore, to enhance the model's capacity to capture TB lesion features, we introduce an improved iterative optimization algorithm that refines feature maps by integrating probability maps obtained in previous iterations. Extensive experiments validate the effectiveness of DPTBNet and the practicality of the DeepPulmoTB dataset.http://www.sciencedirect.com/science/article/pii/S2405844024015214Multi-task learningSegmentationClassificationCT tuberculosis imaging |
spellingShingle | Zhuoyi Tan Hizmawati Madzin Bahari Norafida Yang ChongShuang Wei Sun Tianyu Nie Fengzhou Cai DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT) Heliyon Multi-task learning Segmentation Classification CT tuberculosis imaging |
title | DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT) |
title_full | DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT) |
title_fullStr | DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT) |
title_full_unstemmed | DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT) |
title_short | DeepPulmoTB: A benchmark dataset for multi-task learning of tuberculosis lesions in lung computerized tomography (CT) |
title_sort | deeppulmotb a benchmark dataset for multi task learning of tuberculosis lesions in lung computerized tomography ct |
topic | Multi-task learning Segmentation Classification CT tuberculosis imaging |
url | http://www.sciencedirect.com/science/article/pii/S2405844024015214 |
work_keys_str_mv | AT zhuoyitan deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct AT hizmawatimadzin deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct AT baharinorafida deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct AT yangchongshuang deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct AT weisun deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct AT tianyunie deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct AT fengzhoucai deeppulmotbabenchmarkdatasetformultitasklearningoftuberculosislesionsinlungcomputerizedtomographyct |