IoU-Based Anchor Box Estimation for Enhanced Lung Region Localization in Chest X-rays Using YOLO v4

Öz

Precise lung region detection in chest radiographs is an essential preprocessing step for computer-aided diagnostics. This study presents a YOLO v4–based framework to automatically localize lung regions in posteroanterior (PA) chest X-rays. A subset of the CheXpert dataset, containing 456 manually annotated PA radiographs, was used. Anchor boxes were estimated via an Intersection-over-Union (IoU)–based clustering method, improving scale invariance and shape alignment over Euclidean metrics. Empirical evaluation showed that six anchor boxes achieved the best balance between mean IoU (0.883) and computational efficiency. The trained model was tested on 144 images, yielding Average Precision (AP) of 0.9043 for the lung_region class, which represents only the anatomical lung area and not any specific pathology. The precision–recall curve indicated high precision across most recall values, and the confusion matrix showed 124 true positives, 13 false positives, and 7 false negatives. These results demonstrate that YOLO v4 with optimized anchor box estimation enables accurate, efficient lung region localization, supporting automated radiology workflows.

Anahtar Kelimeler

• Image Processing
• YOLO v4
• Anchor Box Optimization
• Lung Region Detection

YOLO v4 Kullanarak Göğüs Röntgenlerinde Geliştirilmiş Akciğer Bölgesi Yerelleştirme için IoU Tabanlı Çapa Kutusu Tahmini

Öz

Göğüs röntgenlerinde akciğer bölgesinin doğru tespiti, bilgisayar destekli tanı sistemleri için kritik bir ön işleme adımıdır. Bu çalışmada, posteroanterior (PA) göğüs röntgenlerinde akciğer bölgelerini otomatik olarak yerelleştirmek için YOLO v4 tabanlı bir çerçeve sunulmuştur. CheXpert veri kümesinden 456 adet elle anotlanmış PA röntgen kullanılmıştır. Çapa kutuları, ölçekten bağımsız mesafe ölçümü ve şekil hizalamasında iyileşme sağlayan IoU (Intersection-over-Union) tabanlı kümeleme yöntemiyle tahmin edilmiştir. Deneysel değerlendirmeler, altı çapa kutusunun ortalama IoU (0,883) ve hesaplama verimliliği açısından en iyi dengeyi sunduğunu göstermiştir. Eğitilen model, 144 görüntüden oluşan test kümesinde çalıştırılmış ve lung_region sınıfı, yalnızca anatomik akciğer bölgesini temsil etmekte olup herhangi bir patolojiyi göstermemektedir; Doğruluk (AP) değeri 0,9043 elde edilmiştir. Kesinlik–duyarlılık eğrisi, çoğu duyarlılık değerinde yüksek kesinlik göstermiştir. Karmaşıklık matrisi ise 124 doğru pozitif, 13 yanlış pozitif ve 7 yanlış negatif tespit etmiştir. Sonuçlar, optimize edilmiş çapa kutusu tahmini ile YOLO v4’ün doğru ve verimli akciğer bölgesi yerelleştirmesi sağlayabildiğini göstermektedir.

Anahtar Kelimeler

• Akciğer Bölgesi Tespiti
• Görüntü İşleme
• YOLO v4
• Çapa Kutusu Optimizasyonu
• Nesne Tespiti

Kaynakça

1. Abut, S. (2024). AI-based model design for prediction of COPD grade from chest X-ray images: a model proposal (COPD-GradeNet). Cukurova University Journal of the Faculty of Engineering, 39(2), 325-338.
2. Abut, S., & Okut, H. (2024). The Importance of Artificial Neural Networks in Decision Making for the Field of Medicine. In G. A. Indrajit, Mittal; Hemlata, Jain (Ed.), The Future of Artificial Neural Networks (pp. 1-24). Nova Science. https://doi.org/10.52305/YUZX7201
3. Abut, S., Okut, H., & Kallail, K. J. (2024). Paradigm shift from Artificial Neural Networks (ANNs) to deep Convolutional Neural Networks (DCNNs) in the field of medical image processing. Expert Systems with Applications, 244, 122983.
4. Acharya, A. K., & Satapathy, R. (2020). A deep learning based approach towards the automatic diagnosis of pneumonia from chest radio-graphs. Biomedical and Pharmacology Journal, 13(1), 449-455.
5. Ait Nasser, A., & Akhloufi, M. A. (2023). A review of recent advances in deep learning models for chest disease detection using radiography. Diagnostics, 13(1), 159.
6. Ausawalaithong, W., Thirach, A., Marukatat, S., & Wilaiprasitporn, T. (2018, 21-24 Nov. 2018). Automatic Lung Cancer Prediction from Chest X-ray Images Using the Deep Learning Approach. 2018 11th Biomedical Engineering International Conference (BMEiCON),
7. Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934.
8. Çallı, E., Sogancioglu, E., Van Ginneken, B., van Leeuwen, K. G., & Murphy, K. (2021). Deep learning for chest X-ray analysis: A survey. Medical image analysis, 72, 102125.

9. Ergen, B., & Abut, S. (2013). Gender recognition using facial images. International Proceedings of Chemical, Biological & Environmental Engineering, 60(22), 112-117. https://doi.org/10.7763/IPCBEE.2013.V60.22
10. Han, Y., Chen, C., Tewfik, A., Glicksberg, B., Ding, Y., Peng, Y., & Wang, Z. (2022). Knowledge-augmented contrastive learning for abnormality classification and localization in chest X-rays with radiomics using a feedback loop. Proceedings of the IEEE/CVF winter conference on applications of computer vision,
11. Holzinger, A., Saranti, A., Molnar, C., Biecek, P., & Samek, W. (2020). Explainable AI methods-a brief overview. International workshop on extending explainable AI beyond deep models and classifiers,
12. Irvin, J., Rajpurkar, P., Ko, M., Yu, Y., Ciurea-Ilcus, S., Chute, C., Marklund, H., Haghgoo, B., Ball, R., & Shpanskaya, K. (2019). Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison. Proceedings of the AAAI conference on artificial intelligence,
13. Matsuyama, E. (2021). A novel method for automated lung region segmentation in chest X-ray images. Journal of biomedical science and engineering, 14(6), 288-299.
14. Miah, M. A. I., Paul, S., Das, S., & Hashem, M. (2024). Inflocnet: Enhanced lung infection localization and disease detection from chest x-ray images using lightweight deep learning. arXiv preprint arXiv:2408.06459.
15. Nguyen, H. T., Nguyen, M. N., Phung, L. D., & Pham, L. T. T. (2023). Anomalies Detection in Chest X-Rays Images Using Faster R-CNN and YOLO. Vietnam Journal of Computer Science, 10(04), 499-515. https://doi.org/10.1142/s2196888823500094
16. Petsiuk, V., Jain, R., Manjunatha, V., Morariu, V. I., Mehra, A., Ordonez, V., & Saenko, K. (2021). Black-box explanation of object detectors via saliency maps. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition,
17. Ragab, M., Jadid Abdulkadir, S., Muneer, A., Alqushaibi, A., Sumiea, E., Qureshi, R., Al-Selwi, S., & Alhussian, H. (2024). A Comprehensive Systematic Review of YOLO for Medical Object Detection (2018 to 2023). IEEE Access, PP, 1-1. https://doi.org/10.1109/ACCESS.2024.3386826
18. Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D., Bagul, A., Langlotz, C., & Shpanskaya, K. (2017). Chexnet: Radiologist-level pneumonia detection on chest x-rays with deep learning. arXiv preprint arXiv:1711.05225.
19. Redmon, J., & Farhadi, A. (2016). YOLO9000: Better, Faster, Stronger. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 6517-6525.
20. Redmon, J., & Farhadi, A. (2017). YOLO9000: better, faster, stronger. Proceedings of the IEEE conference on computer vision and pattern recognition,
21. Saporta, A., Gui, X., Agrawal, A., Pareek, A., Truong, S. Q. H., Nguyen, C. D. T., Ngo, V.-D., Seekins, J., Blankenberg, F. G., Ng, A. Y., Lungren, M. P., & Rajpurkar, P. (2022). Benchmarking saliency methods for chest X-ray interpretation. Nature Machine Intelligence, 4(10), 867-878. https://doi.org/10.1038/s42256-022-00536-x
22. Yan, S. (2020). Detection Method of Chest X-ray Lung Nodules Based on Improved YOLOV3 Model. Science Discovery, 8(1), 18-23. https://doi.org/10.11648/j.sd.20200801.15