doi: 10.7763/IJCTE.2025.V17.1365
A Comparison of 2D and 3D CNN for Lung CT Image Tuberculosis Severity Assessment
David Olayemi Alebiosu1,*
,
Farrukh Hassan1 ,
Adeola Folayan2 ,
Mathew Yit Hang Yeow1
- 1Department of Computing and Information Systems, Sunway University, Bandar Sunway, Malaysia
- 2Jeffrey Cheah School of Medicine and Health Sciences, Monash University, Malaysia
- Manuscript receivedJuly 29, 2024
- revisedSeptember 9, 2024
- acceptedDecember 3, 2024
- publishedFebruary 14, 2025
Abstract
Tuberculosis is an infectious disease that usually affects the lungs. However, early diagnosis of tuberculosis increases the chance of cure. Practical analysis of lung Computed Tomography (CT) images from tuberculosis patients is one of the primary methods used to determine the severity of the disease. Handcrafted CT image analysis techniques such as grey level concurrence matrix, Fourier transform, etc, used for medical image pre-processing techniques have been ineffective due to their limitations in extracting discriminating features from the images. The application of deep learning, a branch of machine learning, is gaining increased acceptance in medical image analysis. The challenges such as high cost, human error, and slow speed encountered during manual labelling are gradually eliminated in various scales with deep learning techniques. This study explores two deep-learning approaches to classify TB severity in Lung CT Images. Two-Dimensional (2D) and three-Dimensional (3D) convolutional neural networks (CNNs) were used separately to classify the ImageCLEF 2021 lung CT dataset into ‘High’ and ‘Low’ severity categories. The proposed 3D-CNN in this study outperformed the 2D CNNs; it produced an overall average accuracy and Area Under the ROC curve (AUC) of 0.9929 and 0.9982 respectively.
Keywords
- tuberculosis
- lung CT
- classification
How to Cite
David Olayemi Alebiosu, Farrukh Hassan, Adeola Folayan, and Mathew Yit Hang Yeow, "A Comparison of 2D and 3D CNN for Lung CT Image Tuberculosis Severity Assessment," International Journal of Computer Theory and Engineering, vol. 17, no. 1, pp. 21-27, 2025. https://doi.org/10.7763/IJCTE.2025.V17.1365
Copyright & License
Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).