LXNet: A lightweight CNN for lung disease classification from Chest X-ray with XAI-based interpretability
Juiria Humayan, Md. Najmus Sakib Nahid, Amir Sohel, Md Alamgir Kabir, Md Shakhawat Hossain, Zahid Ullah, Mona Jamjoom
Abstract
The diagnosis of lung diseases such as pneumonia and tuberculosis remains a major global health challenge, especially in resource-limited regions. Artificial Intelligence (AI) has shown strong potential in analyzing Chest X-Rays (CXR) for accurate and timely diagnosis, but most existing models are computationally heavy and lack interpretability, limiting their practical application.
Introduction
Lung diseases remain a primary global health concern, contributing significantly to morbidity and mortality worldwide. According to the World Health Organization (WHO), tuberculosis alone accounts for over 1.3 million deaths annually [1], while respiratory infections such as pneumonia and COVID-19 continue to place substantial burdens on healthcare systems.
Methods
This study was systematically designed to enable a comprehensive and rigorous evaluation of the proposed LXNet architecture for multi-class lung disease classification using CXR images. Initially, the lung CXR dataset was balanced and preprocessed to ensure high-quality inputs suitable for model training.
Results
A series of ablation experiments was performed to evaluate the effect of key hyperparameters on LXNet’s performance. Each parameter was varied independently, while the others were kept at their baseline settings.
Discussion
In this study, we have demonstrated the accuracy and validated the lightweight, interpretable properties of LXNet through experiments. The proposed LXNet achieved state-of-the-art performance in multi-class lung disease classification of CXR images. As shown in Table 15, LXNet attained an overall accuracy of 99.62%, surpassing previous models.
Conclusion
This study introduced LXNet, a lightweight and interpretable CNN designed for multi-class lung disease classification from CXR images. Through comprehensive evaluation and comparison with state-of-the-art pretrained models, LXNet achieved high accuracy (macro-average accuracy of 99.62% on the hold-out test and 96.1% under 5-fold cross-validation) and computational efficiency.
Citation: Humayan J, Nahid MNS, Sohel A, Kabir MA, Hossain MS, Ullah Z, et al. (2026) LXNet: A lightweight CNN for lung disease classification from Chest X-ray with XAI-based interpretability. PLoS One 21(6): e0351762. https://doi.org/10.1371/journal.pone.0351762
Editor: Asadullah Shaikh, Najran University College of Computer Science and Information Systems, SAUDI ARABIA
Received: October 22, 2025; Accepted: May 29, 2026; Published: June 17, 2026
Copyright: © 2026 Humayan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The dataset used in this study is publicly available at: https://www.kaggle.com/datasets/fernando2rad/x-ray-lung-diseases-images-9-classes. The code developed for this study is publicly available at: https://github.com/sakibnjr/LXNet-----Lightweight-CNN/tree/main/Updated.
Funding: This work was supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2026R104), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.
Competing interests: NO authors have competing interests.
