Derin Öğrenme Yaklaşımları Kullanarak Göğüs Röntgeni Görüntülerinden Atelektazinin Sınıflandırılması

Abstract

Atelektazi, akciğerin kısmen veya tamamen sönmesiyle karakterize edilen ve erken tanı gerektiren önemli bir solunum sistemi hastalığıdır. Bu çalışmada, göğüs röntgeni görüntülerinden atelektazinin otomatik sınıflandırılmasında İlgi Bölgesi (ROI) kırpma ve transfer öğrenmesinin etkileri incelenmiştir. SqueezeNet mimarisi kullanılarak dört farklı model geliştirilmiştir. Modeller, iki görüntü türü (ham ve akciğer bölgesi kırpılmış görüntüler) ile iki eğitim stratejisine (sıfırdan eğitim ve transfer öğrenmesi) göre tasarlanmıştır. Model I ve II sırasıyla ham ve kırpılmış görüntüler üzerinde sıfırdan eğitilmiş, Model III ve IV ise aynı görüntü türlerinde transfer öğrenmesi kullanılarak geliştirilmiştir. Performans standart sınıflandırma ölçütleriyle değerlendirilmiştir. En yüksek performans, transfer öğrenmesi ve ROI kırpma yöntemini birleştiren Model IV tarafından elde edilmiştir. Model %98,43 doğruluk, 0,984 F1 skoru ve 0,999 AUC değeri sağlamıştır. ROI kırpmanın her iki eğitim stratejisinde de performansı artırdığı gözlenmiştir. Transfer öğrenmesi ile ROI tabanlı ön işlemenin birlikte kullanılması, atelektazi tespitinde yüksek başarı sağlayarak klinik karar destek sistemleri için umut verici sonuçlar sunmaktadır.

Keywords

• Atelektazi
• Derin Öğrenme
• Transfer Öğrenme
• X-ışını Görüntüleme
• SqueezeNet

Supporting Institution

TUBITAK

Project Number

1059B192302217

Ethical Statement

Bu makale etik standartlara uygundur.

Thanks

F.D., Türkiye Bilim ve Teknolojik Araştırma Kurumu'na (TUBITAK, 2219 - Uluslararası Doktora Sonrası Araştırma Burs Programı, 1059B192302217) aldığı doktora sonrası araştırma bursu nedeniyle teşekkür etmektedir.

Classification of Atelectasis from Chest X-Ray Images Using Deep Learning Approaches

Abstract

Atelectasis, the partial or complete collapse of the lung, requires early diagnosis for effective treatment. This study investigates the effects of Region of Interest (ROI) cropping and transfer learning on the automatic classification of atelectasis from chest X-ray images. Four SqueezeNet-based models were developed using two image types (raw and cropped lung regions) and two training strategies (from-scratch and transfer learning). Models I and II were trained from scratch on raw and cropped images, respectively, while Models III and IV employed transfer learning on the same image sets. Performance was evaluated using standard classification metrics. Model IV, combining transfer learning and ROI-cropped images, achieved the best performance, with 98.43% accuracy, 0.984 F1-score, and 0.999 AUC. It also reached 0.9861 precision, 0.9828 sensitivity, and 0.9859 specificity. ROI-based cropping consistently improved classification performance. Combining transfer learning with ROI-focused preprocessing significantly enhances atelectasis detection and shows promise as a reliable clinical decision-support approach.

Keywords

• Atelectasis
• Deep Learning
• Transfer Learning
• X-ray Imaging
• SqueezeNet
• Image Classification

Supporting Institution

the Scientific and Technical Research Council of Turkey

Project Number

1059B192302217

Ethical Statement

This article complies with ethical standards

Thanks

F.D. acknowledges a Postdoctoral grant from the Scientific and Technical Research Council of Turkey (TUBITAK, 2219 - International Postdoctoral Research Scholarship Programme, 1059B192302217).

References

1. Peroni, D.G. and A.L. Boner, Atelectasis: mechanisms, diagnosis and management. Paediatric Respiratory Reviews, 2000. 1(3): p. 274-278.
2. Woodring, J.H. and J.C. Reed, Types and Mechanisms of Pulmonary Atelectasis. Journal of Thoracic Imaging, 1996. 11(2): p. 92-108.
3. Berikol, G.B., et al., Mapping artificial intelligence models in emergency medicine: A scoping review on artificial intelligence performance in emergency care and education. Turkish Journal of Emergency Medicine, 2025. 25(2): p. 67-91.
4. Carlsen, K.H., S. Crowley, and B. Smevik, 70 - Atelectasis, in Kendig's Disorders of the Respiratory Tract in Children (Ninth Edition), R.W. Wilmott, et al., Editors. 2019, Elsevier: Philadelphia. p. 1027-1033.e1.
5. Restrepo, R.D. and J. Braverman, Current challenges in the recognition, prevention and treatment of perioperative pulmonary atelectasis. Expert Review of Respiratory Medicine, 2015. 9(1): p. 97-107.
6. Hedenstierna, G. and H.U. Rothen, Atelectasis Formation During Anesthesia: Causes and Measures to Prevent It. Journal of Clinical Monitoring and Computing, 2000. 16(5): p. 329-335.
7. Sajed, S., et al., The effectiveness of deep learning vs. traditional methods for lung disease diagnosis using chest X-ray images: A systematic review. Applied Soft Computing, 2023. 147: p. 110817.
8. Ayalew, A.M., et al., Atelectasis detection in chest X-ray images using convolutional neural networks and transfer learning with anisotropic diffusion filter. Informatics in Medicine Unlocked, 2024. 45: p. 101448.

9. Abut, S., H. Okut, and K.J. Kallail, Paradigm shift from Artificial Neural Networks (ANNs) to deep Convolutional Neural Networks (DCNNs) in the field of medical image processing. Expert Systems with Applications, 2024. 244: p. 122983.
10. Hwang, E.J., et al., Deep Learning for Detection of Pulmonary Metastasis on Chest Radiographs. Radiology, 2021. 301(2): p. 455-463.
11. Moses, D.A., Deep learning applied to automatic disease detection using chest X-rays. Journal of Medical Imaging and Radiation Oncology, 2021. 65(5): p. 498-517.
12. Abut, S., et al., Deep Neural Networks and Applications in Medical Research, in Deep Learning - Recent Findings and Research, M. Domínguez-Morales, et al., Editors. 2023, IntechOpen: London.
13. Yang, S., et al., Intelligent health care: Applications of deep learning in computational medicine. Frontiers in Genetics, 2021. 12: p. 607471.
14. Ching, T., et al., Opportunities and obstacles for deep learning in biology and medicine. Journal of The Royal Society Interface, 2018. 15(141): p. 20170387.
15. Bakator, M. and D. Radosav, Deep Learning and Medical Diagnosis: A Review of Literature. Multimodal Technologies and Interaction, 2018. 2(3): p. 47.
16. Ergen, B. and S. Abut. Gender recognition using facial images.
17. Jasmine Pemeena Priyadarsini, M., et al., Lung Diseases Detection Using Various Deep Learning Algorithms. Journal of Healthcare Engineering, 2023. 2023(1): p. 3563696.
18. Kido, S., Y. Hirano, and N. Hashimoto. Detection and classification of lung abnormalities by use of convolutional neural network (CNN) and regions with CNN features (R-CNN). in 2018 International Workshop on Advanced Image Technology (IWAIT). 2018.
19. Abut, S., AI-Based Model Design for Prediction of COPD Grade from Chest X-Ray Images: A Model Proposal (COPD-GradeNet). Çukurova Üniversitesi Mühendislik Fakültesi Dergisi, 2024. 39(2): p. 325-338.
20. Aggarwal, R., et al., Diagnostic accuracy of deep learning in medical imaging: a systematic review and meta-analysis. npj Digital Medicine, 2021. 4(1): p. 65.
21. Geng, Y., et al. 2.5 D CNN model for detecting lung disease using weak supervision. in Medical imaging 2019: computer-aided diagnosis. 2019. SPIE.
22. Tharakan, J.J., M. Kanchan, and O.S. Powar. Enhancing Disease Detection with Machine and Deep Learning: Analyzing Chest X-rays for Pneumonia and Atelectasis Identification. in 2024 International Conference on Augmented Reality, Intelligent Systems, and Industrial Automation (ARIIA). 2024.
23. Nam, J.G., et al., Development and validation of a deep learning algorithm detecting 10 common abnormalities on chest radiographs. European Respiratory Journal, 2021. 57(5).
24. Rajpurkar, P., et al., Chexnet: Radiologist-level pneumonia detection on chest x-rays with deep learning. arXiv preprint arXiv:1711.05225, 2017.
25. Chutia, U., A.S. Tewari, and J.P. Singh, Collapsed lung disease classification by coupling denoising algorithms and deep learning techniques. Network Modeling Analysis in Health Informatics and Bioinformatics, 2023. 13(1): p. 1.
26. Chutia, U., et al., Classification of lung diseases using an attention-based modified DenseNet model. Journal of Imaging Informatics in Medicine, 2024. 37(4): p. 1625-1641.
27. Chutia, U., A.S. Tewari, and J.P. Singh, Lung disease classification using deep learning and genetic algorithm. Neural Computing and Applications, 2024.
28. Abut, S.O., Hayrettin, The Importance of Artificial Neural Networks in Decision Making for the Field of Medicine, in The Future of Artificial Neural Networks, I.M. Ghosal, Arun; Jain, Hemlata, Editor. 2024, Nova Science and Technology: New York. p. 1-24.
29. Oflas, R.D., Fatih; Abut, Serdar. Anatomical Landmark Detection in Medical Imaging: Deep Learning–Based Approaches and Clinical Reliability in 4th International Conference on Pioneer and Innovative Studies ICPIS 2026. 2026. Konya: All Sciences Academy.
30. Abut, S., IoU-Based Anchor Box Estimation for Enhanced Lung Region Localization in Chest X-rays Using YOLO v4. Kirklareli University Journal of Engineering and Science, 2025. 11(2): p. 253-266.
31. Chaudhary, A., A. Hazra, and P. Chaudhary. Diagnosis of chest diseases in x-ray images using deep convolutional neural network. in 2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT). 2019. IEEE.
32. Saporta, A., et al., Benchmarking saliency methods for chest X-ray interpretation. Nature Machine Intelligence, 2022. 4(10): p. 867-878.
33. Iandola, F.N., et al., SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size. arXiv preprint arXiv:1602.07360, 2016.
34. Tsivgoulis, M., T. Papastergiou, and V. Megalooikonomou, An improved SqueezeNet model for the diagnosis of lung cancer in CT scans. Machine Learning with Applications, 2022. 10: p. 100399.
35. Hassanpour, M. and H. Malek. Document Image Classification using SqueezeNet Convolutional Neural Network. in 2019 5th Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS). 2019.
36. Ullah, A., et al., Comparative analysis of AlexNet, ResNet18 and SqueezeNet with diverse modification and arduous implementation. Arabian journal for science and engineering, 2022. 47(2): p. 2397-2417.
37. Pradeep, K., et al. EdgeNet: SqueezeNet like Convolution Neural Network on Embedded FPGA. in 2018 25th IEEE International Conference on Electronics, Circuits and Systems (ICECS). 2018.
38. Yuan, C., et al., An effective image classification method for shallow densely connected convolution networks through squeezing and splitting techniques. Applied Intelligence, 2019. 49(10): p. 3570-3586.
39. Kingma, D.P. and J. Ba, Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.
40. Hospodarskyy, O., et al., Understanding the adam optimization algorithm in machine learning. 2024, 2024.
41. Smaida, M., S. Yaroshchak, and A.B. Sasi, Learning rate optimization in CNN for accurate ophthalmic classification. Int. J. Innov. Technol. Explor. Eng.(IJITEE), 2021. 10: p. 211-216.
42. Masters, D. and C. Luschi, Revisiting small batch training for deep neural networks. arXiv preprint arXiv:1804.07612, 2018.
43. Rubinstein, R., The cross-entropy method for combinatorial and continuous optimization. Methodology and computing in applied probability, 1999. 1(2): p. 127-190.
44. Botev, Z.I., et al., The cross-entropy method for optimization, Handbook of Statistics. Handbook of Statistics. Machine Learning, 2013. 31: p. 19-34.
45. De Boer, P.-T., et al., A tutorial on the cross-entropy method. Annals of operations research, 2005. 134(1): p. 19-67.
46. Sharma, A., Testing of machine learning algorithms and models. 2023, Universität Oldenburg.
47. Botchkarev, A., A new typology design of performance metrics to measure errors in machine learning regression algorithms. Interdisciplinary Journal of Information, Knowledge, and Management, 2019. 14: p. 045-076.
48. Naidu, G., T. Zuva, and E.M. Sibanda. A Review of Evaluation Metrics in Machine Learning Algorithms. in Artificial Intelligence Application in Networks and Systems. 2023. Cham: Springer International Publishing.
49. Wu, S. and P. Flach. A scored AUC metric for classifier evaluation and selection. in Second workshop on ROC analysis in ML, bonn, Germany. 2005.
50. Zhang, X., et al., The use of ROC and AUC in the validation of objective image fusion evaluation metrics. Signal processing, 2015. 115: p. 38-48.
51. Muschelli III, J., ROC and AUC with a binary predictor: a potentially misleading metric. Journal of classification, 2020. 37(3): p. 696-708.
52. Marzban, C., The ROC curve and the area under it as performance measures. Weather and Forecasting, 2004. 19(6): p. 1106-1114.
53. Ronneberger, O., P. Fischer, and T. Brox. U-net: Convolutional networks for biomedical image segmentation. in International Conference on Medical image computing and computer-assisted intervention. 2015. Springer.
54. He, K., et al. Mask r-cnn. in Proceedings of the IEEE international conference on computer vision. 2017.
55. Chen, L.-C., et al., Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587, 2017.
56. Oktay, O., et al., Attention u-net: Learning where to look for the pancreas. arXiv preprint arXiv:1804.03999, 2018.