International Journal For Multidisciplinary Research

E-ISSN: 2582-2160     Impact Factor: 9.24

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Call for Paper Volume 7, Issue 1 (January-February 2025) Submit your research before last 3 days of February to publish your research paper in the issue of January-February.

Deep Learning-Based Classification of Pulmonary Diseases from Chest X-Ray Images

Author(s) Burcu Oltu, Berna Dengiz, Selda Güney
Country Turkey
Abstract COVID-19, tuberculosis (TB), and pneumonia, are life-threatening diseases, that lead to death if not diagnosed or treated promptly. Chest X-rays (CXRs) are the primary imaging technique for detecting pulmonary diseases due to their ease of use, accessibility, low radiation doses, and reasonable pricing. However, interpreting CXRs is highly dependent on radiologists’ experience and is prone to diagnostic errors. Therefore, a computer-aided diagnostic system may improve the detection accuracy. In this study, a deep learning-based (DL) end-to-end model is proposed for the classification of CXRs into TB, pneumonia, COVID-19, lung opacity, and healthy classes. The proposed model uses DenseNet201 as the backbone model for feature extraction and includes a squeeze-and-excitation block, and global average pooling (GAP) to highlight representative features while suppressing the redundant ones. This approach results in an average test accuracy of 98.94%, precision of 98.22%, recall of 98.03%, specificity of 99.18%, F1-score of 98.12%, and area under curve (AUC) of 0.996 for classifying five categories. This DL-based model provides objective results and performs better than existing methods in literature. Ablation studies are also conducted to show the effectiveness of the proposed method. Additionally, Grad-CAM is used to provide a visual representation of the model's decision-making process.
Keywords Chest X-Ray, Deep Learning, Convolutional Neural Networks, Attention Mechanism, Grad-CAM
Published In Volume 7, Issue 1, January-February 2025
Published On 2025-01-16
Cite This Deep Learning-Based Classification of Pulmonary Diseases from Chest X-Ray Images - Burcu Oltu, Berna Dengiz, Selda Güney - IJFMR Volume 7, Issue 1, January-February 2025. DOI 10.36948/ijfmr.2025.v07i01.33673
DOI https://doi.org/10.36948/ijfmr.2025.v07i01.33673
Short DOI https://doi.org/g82hkm

Share this