Conference Paper2025

Deep Learning for Accurate Classification of Pulmonary Edema Severity in Chest X-Rays: A Comparative Evaluation of CNN Models with Grad-CAM Feature Visualization

Manosh Sur Choudhury, Rashedur Rahman, Siam Tahsin Bhuiyan, Sefatul Wasi, Saadia Binte Alam

2025 International Conference on Wavelet Analysis and Pattern Recognition (ICWAPR)

IEEE, pp. 72-77

  • Good Health and Well-being — Relevance: 95%
DOI: 10.1109/icwapr66259.2025.11306863URL

Abstract

Accurate classification of pulmonary edema severity is essential for timely diagnosis and effective management, as each severity level requires distinct therapeutic interventions. Differentiating between edema severity classes is particularly challenging due to the overlapping radiographic features of Chest X-ray (CXR). Deep learning, especially convolutional neural networks (CNNs), presents a promising solution by automating classification and identifying subtle features often difficult to detect through conventional methods. This study evaluates the performance of six traditional deep learning models for classifying pulmonary edema severity. Among these, CheXNet achieved the best performance, with an accuracy of 91% and an overall AUC score of 0.88. The findings highlight the importance of using pretraining on CXR datasets, which significantly enhances model performance compared to general pretraining. Additionally, Grad-CAM visualization was employed to interpret model decisions, identifying key radiographic features that contribute to accurate classification across different severity levels.

Keywords

  • convolutional neural networks
  • Grad-CAM
  • CheXNet
  • chest radiography
  • Pulmonary edema
  • transfer learning
  • model interpretability

CCDS Authors

Manosh Sur Choudhury
Manosh Sur Choudhury

Manosh Sur Choudhury

  • Research Assistant, CCDS

Masters

+1 more affiliation

Dhaka University

Md. Rashedur Rahman, D. Eng.
Md. Rashedur Rahman, D. Eng.

Md. Rashedur Rahman, D. Eng.

  • Co-Director, MIRA & NEST Wings

Assistant Professor

Department of Computer Science and Engineering

Independent University, Bangladesh

Published at

CHI, HCII, Scientific Reports, Cognitive Computation, Neural Computing and Applications

Siam Tahsin Bhuiyan
Siam Tahsin Bhuiyan

Siam Tahsin Bhuiyan

  • Intern, CCDS

Bachelors

Independent University, Bangladesh

Published at

Scientific Reports

Saadia Binte Alam, PhD
Saadia Binte Alam, PhD

Saadia Binte Alam, PhD

  • Director, MIRA & WiSE Wings

Associate Professor

Department of Computer Science and Engineering

Independent University, Bangladesh

Published at

HCII, Scientific Reports, Cognitive Computation, Neural Computing and Applications

References

  1. 1.Irvin, Jeremy, Rajpurkar, Pranav, Ko, Michael, Yu, Yifan, Ciurea-Ilcus, Silviana, Chute, Chris, Marklund, Henrik, Haghgoo, Behzad, Ball, Robyn, Shpanskaya, Katie, Seekins, Jayne, Mong, David A., Halabi, Safwan S., Sandberg, Jesse K., Jones, Ricky, Larson, David B., Langlotz, Curtis P., Patel, Bhavik N., Lungren, Matthew P., Ng, Andrew Y.. (2019). CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparison. AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI))[10.1609/aaai.v33i01.3301590]
  2. 2.Pranav Rajpurkar, Jeremy Irvin, Robyn L. Ball, Kaylie Zhu, Brandon Yang, Hershel Mehta, Tony Duan, Daisy Ding, Aarti Bagul, Curtis P. Langlotz, Bhavik N. Patel, Kristen W. Yeom, Katie Shpanskaya, Francis G. Blankenberg, Jayne Seekins, Timothy J. Amrhein, David A. Mong, Safwan S. Halabi, Evan J. Zucker, Andrew Y. Ng, Matthew P. Lungren. (2018). Deep learning for chest radiograph diagnosis: A retrospective comparison of the CheXNeXt algorithm to practicing radiologists. PLoS Medicine, 15(11), e1002686[10.1371/journal.pmed.1002686]
  3. 3.Goram Mufarah M. Alshmrani, Qiang Ni, Richard Jiang, Haris Pervaiz, Nada M. Elshennawy. (2022). A deep learning architecture for multi-class lung diseases classification using chest X-ray (CXR) images. Alexandria Engineering Journal, 64, 923–935[10.1016/j.aej.2022.10.053]
  4. 4.Vinayakumar Ravi, Vasundhara Acharya, Mamoun Alazab. (2022). A multichannel EfficientNet deep learning-based stacking ensemble approach for lung disease detection using chest X-ray images. Cluster Computing, 26(2), 1181–1203[10.1007/s10586-022-03664-6]
  5. 5.Eunbee Kim, Anthony J. Fischetti, Pratheev Sreetharan, Joel G. Weltman, Philip R. Fox. (2022). Comparison of artificial intelligence to the veterinary radiologist's diagnosis of canine cardiogenic pulmonary edema. Veterinary Radiology & Ultrasound, 63(3), 292–297[10.1111/vru.13062]
  6. 6.AliNawaz Khan, Hamdan AL-Jahdali, S AL-Ghanem, Alaa Gouda. (2009). Reading chest radiographs in the critically ill (Part II): Radiography of lung pathologies common in the ICU patient. Annals of Thoracic Medicine, 4(3), 149[10.4103/1817-1737.53349]
  7. 7.Steven Horng, Ruizhi Liao, Xin Wang, Sandeep Dalal, Polina Golland, Seth J. Berkowitz. (2021). Deep Learning to Quantify Pulmonary Edema in Chest Radiographs. Radiology Artificial Intelligence, 3(2), e190228[10.1148/ryai.2021190228]
  8. 8.Qinghua Huang, Ye Lei, Wenyu Xing, Chao He, Gaofeng Wei, Zhaoji Miao, Yifan Hao, Guannan Li, Yan Wang, Qingli Li, Xuelong Li, Wenfang Li, Jiangang Chen. (2022). Evaluation of Pulmonary Edema Using Ultrasound Imaging in Patients With COVID-19 Pneumonia Based on a Non-local Channel Attention ResNet. Ultrasound in Medicine & Biology, 48(5), 945–953[10.1016/j.ultrasmedbio.2022.01.023]
  9. 9.Viacheslav V Danilov, Anton O Makoveev, Alex Proutski, Irina Ryndova, Alex Karpovsky, Yuriy Gankin. (2024). Explainable AI to identify radiographic features of pulmonary edema. Radiology Advances, 1(1), umae003[10.1093/radadv/umae003]
← All publications