DERİN ÖĞRENME MODELLERİ KULLANARAK RETİNA HASTALIĞI SINIFLANDIRMASI

Öz

Bu çalışma, derin öğrenme yöntemlerini kullanarak retina hastalıklarının otomatik sınıflandırılmasını ve teşhis süreçlerini iyileştirmeyi amaçlamaktadır. Retina hastalıkları, özellikle diabetik retinopati, yaşa bağlı maküler dejenerasyon (AMD), glokom ve retina damar tıkanıklığı, dünya genelinde görme kaybının başlıca nedenleri arasındadır. Bu hastalıkların erken teşhisi ve doğru sınıflandırılması, görme kaybını önlemek adına kritik bir öneme sahiptir. Derin öğrenme tabanlı yaklaşımlar, insan faktörüne bağlı teşhis hatalarını azaltarak daha yüksek doğruluk oranları sunmakta ve retina görüntüleme yöntemlerinin etkinliğini artırmaktadır. Çalışmada, konvolüsyonel sinir ağları (CNN) ve transfer learning modelleri kullanılarak retina hastalıklarının sınıflandırılmasının karşılaştırılması gerçekleştirilmiştir. Fundus ve optik koherens tomografi (OCT) görüntüleri üzerinde yapılan analizler, yüksek doğruluk oranlarıyla bu yöntemlerin etkili olduğunu ortaya koymaktadır. Elde edilen modeller, doğruluk, hassasiyet, hatırlama ve F1 skoru gibi metriklerle değerlendirilmiş ve klinik uygulamalardaki potansiyelleri irdelenmiştir. Araştırma sonuçları, derin öğrenme yöntemlerinin, retina hastalıklarının erken teşhisinde hız, doğruluk ve tekrarlanabilirlik gibi avantajlar sunduğunu göstermektedir. Özellikle CNN tabanlı modellerin performansı, uzman teşhis süreçlerini destekleyerek görme kaybını önlemeye yönelik önemli bir katkı sağlamaktadır. Bu çalışma, tıbbi görüntüleme teknolojilerinde derin öğrenmenin kullanımına dair yeni bir perspektif sunmakta ve sağlık profesyonellerinin iş yükünü azaltacak çözüm önerileri ortaya koymaktadır.

Anahtar Kelimeler

• derin öğrenme
• retina hastalıkları
• fundus görüntüleme
• transfer learning
• CNN

ENGLISH RETINAL DISEASE CLASSIFICATION USING DEEP LEARNING MODELS

Öz

This study aims to improve the automatic classification and diagnosis processes of retinal diseases using deep learning methods. Retinal diseases, especially diabetic retinopathy, age-related macular degeneration (AMD), glaucoma and retinal vascular occlusion, are among the main causes of vision loss worldwide. Early diagnosis and correct classification of these diseases are of critical importance in preventing vision loss. Deep learning-based approaches offer higher accuracy rates by reducing diagnostic errors due to human factors and increase the efficiency of retinal imaging methods. In the study, a comparison of the classification of retinal diseases was performed using convolutional neural networks (CNN) and transfer learning models. Analyses performed on fundus and optical coherence tomography (OCT) images reveal that these methods are effective with high accuracy rates. The obtained models were evaluated with metrics such as accuracy, precision, recall and F1 score, and their potential in clinical applications was examined. The research results show that deep learning methods offer advantages such as speed, accuracy and reproducibility in the early diagnosis of retinal diseases. In particular, the performance of CNN-based models provides a significant contribution to preventing vision loss by supporting expert diagnosis processes. This study offers a new perspective on the use of deep learning in medical imaging technologies and suggests solutions that will reduce the workload of healthcare professionals.

Anahtar Kelimeler

• deep learning
• retinal diseases
• fundus imaging
• transfer learning
• CNN

Kaynakça

1. Brown, G. C., et al. (2019). Retina. Elsevier.
2. Deye, J., et al. (2020). Multimodal deep learning for retinal disease diagnosis using fundus images and OCT. IEEE Transactions on Medical Imaging, 39(6), 2060–2072. https://doi.org/10.1109/TMI.2019.2952127
3. Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017). Dermatologistlevel classification of skin cancer with deep neural networks. Nature, 542(7639), 115–118. https://doi.org/10.1038/nature21056
4. Gulshan, V., Peng, L., Coram, M., Stumpe, M. C., Wu, D., Narayanaswamy, A., ... & Webster, D. R. (2016). Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA, 316(22), 2402–2410. https://doi.org/10.1001/jama.2016.17216
5. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 770–778). https://doi.org/10.1109/CVPR.2016.90
6. Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017). Densely connected convolutional networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https://doi.org/10.1109/CVPR.2017.243
7. Lee, C. H., et al. (2017). Deep learning in retinal disease diagnosis: A review of current applications. Ophthalmology, 124(4), 593–604. https://doi.org/10.1016/j.ophtha.2016.10.048
8. Leibig, C., Allken, V., Ayhan, M. S., Berens, P., & Wahl, S. (2017). Leveraging uncertainty information from deep neural networks for disease detection. Scientific Reports, 7(1), 1–14. https://doi.org/10.1038/s41598-017-17876-z

9. Litjens, G., Kooi, T., Bejnordi, B. E., Setio, A. A. A., Ciompi, F., Ghafoorian, M., ... & van der Laak, J. (2017). A survey on deep learning in medical image analysis. Medical Image Analysis, 42, 60–88. https://doi.org/10.1016/j.media.2017.07.005
10. Pan, S. J., & Yang, Q. (2010). A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359. https://doi.org/10.1109/TKDE.2009.191
11. RIADD (2021). Retinal Fundus Multi-disease Image Dataset. Erişim adresi: https://riadd.grandchallenge. org/download-all-classes/
12. Ronneberger, O., et al. (2015). U-Net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention (pp. 234–241). https://doi.org/10.1007/978-3-319-24574-4_28
13. Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint. https://arxiv.org/abs/1409.1556
14. Tan, M., & Le, Q. V. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the International Conference on Machine Learning (ICML). https://proceedings.mlr.press/v97/tan19a.html
15. Ting, D. S. W., Pasquale, L. R., Peng, L., Campbell, J. P., Lee, A. Y., Raman, R., ... & Wong, T. Y. (2017). Artificial intelligence and deep learning in ophthalmology. The British Journal of Ophthalmology, 103(2), 167-175. https://doi.org/10.1136/bjophthalmol-2018-313173
16. Ting, D. S. W., Pasquale, L. R., Peng, L., Campbell, J. P., Lee, A. Y., Raman, R., ... & Schmetterer, L. (2017). Artificial intelligence and deep learning in ophthalmology. The British Journal of Ophthalmology, 101(2), 184-190. https://doi.org/10.1136/bjophthalmol-2017-311307
17. Wong, T. Y., et al. (2020). Retinal Imaging. Springer.
18. World Health Organization. (2019). World report on vision. WHO Press. Retrieved from https://www.who.int/publications-detail/world-report-on-vision
19. Zhang, J., Liu, Y., Mitsuhashi, T., & Matsuo, T. (2021). Accuracy of Deep Learning Algorithms for the Diagnosis of Retinopathy of Prematurity by Fundus Images: A Systematic Review and Meta-Analysis. Journal of ophthalmology, 2021, 8883946. https://doi.org/10.1155/2021/8883946
20. Zoph, B., Vasudevan, V., Shlens, J., & Le, Q. V. (2018). Learning transferable architectures for scalable image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https://doi.org/10.1109/CVPR.2018.00916