Incremental learning-based cascaded model for detection and localization of tuberculosis from chest x-ray images

Rapid treatment protocols such as X-ray and CT scans have played a crucial role in the diagnosis of tuberculosis (TB infection). Automatic detection of CXR is required to speed up patient treatment with accuracy. Consequently, it reduces the burden of patients on medical practitioners. The present p...

Full description

Saved in:
Bibliographic Details
Published inExpert systems with applications Vol. 238; p. 122129
Main Authors Vats, Satvik, Sharma, Vikrant, Singh, Karan, Katti, Anvesha, Mohd Ariffin, Mazeyanti, Nazir Ahmad, Mohammad, Ahmadian, Ali, Salahshour, Soheil
Format Journal Article
LanguageEnglish
Published Elsevier Ltd 15.03.2024
Subjects
Cascaded Model
Chest X-Ray
FRCNN
Incremental learning
Radiology
Tuberculosis (TB)
Tuberculosis (TB)
Radiology
Cascaded Model
Incremental learning
Chest X-Ray
FRCNN
Online AccessGet full text

Cover

More Information
Summary:Rapid treatment protocols such as X-ray and CT scans have played a crucial role in the diagnosis of tuberculosis (TB infection). Automatic detection of CXR is required to speed up patient treatment with accuracy. Consequently, it reduces the burden of patients on medical practitioners. The present paper proposes an incremental learning-based cascaded (ILCM) model to detect tuberculosis from Chest X-ray images. The proposed model also localizes the infected region on the CXR image. The experimental outcome, clearly indicates that the performance is better than the pre-trained model as tested on the local population data (93.20% overall accuracy), F1 score of 97.23% (harmonic mean of precision and recall). Where the Golden standard dataset was 83.32% overall accuracy, and F1 score 82.24%.
ISSN:0957-4174
1873-6793
DOI:10.1016/j.eswa.2023.122129