Utilization of Deep Learning Technologies in The Detection and Grading of Diabetic Retinopathy

Yıl 2025, Cilt: 15 Sayı: 4, 1426 - 1446, 15.12.2025

Esra Odabaş Yıldırım , Esma Fazilet Karagülle , Mustafa Yildirim

https://doi.org/10.31466/kfbd.1535708

Öz

Diabetic Retinopathy (DR) is a serious and common complication of diabetes affecting the retinal vessels. As the disease progresses, microaneurysms, hemorrhages, and edema develop in the retinal tissue. If left untreated, it can lead to vision loss and blindness in advanced stages. Therefore, early detection and diagnosis of DR is of great clinical importance. In this study, we evaluated the comparative performance of different Convolutional Neural Network (CCN) and Transformer-based deep learning architectures using retinal images obtained from the Kaggle APTOS 2019 Blindness Detection Competition. The performance of the models is examined in detail in the context of the effects of the applied advanced preprocessing and data augmentation strategies. ESA-based architectures such as ResNet18, EfficientNetB4, VGG16, and DenseNet121, and Transformer-based models such as Swin Transformer were used in the experiments. All models were trained using 5-fold cross-validation, and their performances were compared. The findings revealed that data augmentation techniques significantly increased the accuracy of DR classification. Furthermore, the Swin Transformer model exhibited the highest performance with 85.00% accuracy and 91.37% QWK (Quadratic Weighted Kappa), and Transformer-based models performed better in DR classification compared to traditional ESA-based models.

Anahtar Kelimeler

Convolutional neural networks , Data augmentation , Deep learning , Diabetic retinopathy , Image processing , Transformer architectures

Destekleyen Kurum

Scientific Research Projects

Diyabetik Retinopatinin Tespitinde ve Derecelendirilmesinde Derin Öğrenme Teknolojilerinin Kullanılması

Öz

Diyabetik Retinopati (DR), diyabet hastalığının retina damarlarını etkileyen ciddi ve yaygın bir komplikasyondur. Hastalığın ilerlemesiyle retina dokusunda mikro anevrizmalar, kanamalar ve ödem gelişmekte; tedavi edilmediği takdirde ileri evrelerde görme kaybı ve körlüğe yol açabilmektedir. Bu nedenle DR’nin erken evrede tespiti ve teşhisi klinik açıdan büyük önem taşımaktadır. Bu çalışmada, Kaggle APTOS 2019 Körlük Tespiti yarışmasından elde edilen retina görüntüleri kullanılarak farklı Evrişimli Sinir Ağı (ESA) ve Transformer (Dönüşüm) tabanlı derin öğrenme mimarilerinin karşılaştırmalı olarak performansı değerlendirilmiştir. Modellerin performansları, uygulanan gelişmiş ön işleme ve veri artırma stratejilerinin etkileri bağlamında detaylı olarak incelenmiştir. Deneylerde ResNet18, EfficientNetB4, VGG16 ve DenseNet121 gibi ESA tabanlı mimariler ile Swin Transformer gibi Transformer tabanlı modeller kullanılmıştır. Tüm modeller, 5-katlı çapraz doğrulama yöntemi ile eğitilmiş ve performansları karşılaştırılmıştır. Elde edilen bulgular, veri artırma tekniklerinin DR sınıflandırma başarısını istatistiksel olarak anlamlı düzeyde artırdığını ortaya koymuştur. Ayrıca, Swin Transformer modeli %85.00 doğruluk ve %91.37 QWK (Quadratic Weighted Kappa) ile en yüksek performansı sergilemiş ve Transformer tabanlı modellerin geleneksel ESA tabanlı modellere kıyasla DR sınıflandırmasında daha iyi performans sergilediğini göstermiştir.

Anahtar Kelimeler

Derin öğrenme , Diyabetik retinopati , Evrişimli sinir ağları , Görüntü işleme , Transformer tabanlı mimariler , Veri artırma

Destekleyen Kurum

BAP

Kaynakça

• Akhtar, S., Aftab, S., Ali, O., Ahmad, M., Khan, M. A., Abbas, S., & Ghazal, T. M. (2025). A deep learning based model for diabetic retinopathy grading. Scientific Reports, 15(1), 3763.
• Antonetti, D. A., Klein, R., & Gardner, T. W. (2012). Diabetic retinopathy. New England Journal of Medicine, 366(13), 1227–1239. https://doi.org/10.1056/NEJMra1005073
• Arora, L., Singh, S. K., Kumar, S., Gupta, H., Alhalabi, W., Arya, V., ... & Gupta, B. B. (2024). Ensemble deep learning and EfficientNet for accurate diagnosis of diabetic retinopathy. Scientific Reports, 14(1), 30554.
• Batool, S., Gilani, S. O., Waris, A., Iqbal, K. F., Khan, N. B., Khan, M. I., ... & Awwad, F. A. (2023). Deploying efficient net batch normalizations (BNs) for grading diabetic retinopathy severity levels from fundus images. Scientific Reports, 13(1), 14462.
• Ben-David, A. (2008). Comparison of classification accuracy using Cohen’s Weighted Kappa. Expert Systems with Applications, 34(2), 825-832.
• Bodapati, J. D., Shaik, N. S., & Naralasetti, V. (2021). Composite deep neural network with gated-attention mechanism for diabetic retinopathy severity classification. Journal of Ambient Intelligence and Humanized Computing, 12(10), 9825-9839.
• Boulaabi, M., Gader, T. B. A., Echi, A. K., & Bouraoui, Z. (2025). Enhancing DR Classification with Swin Transformer and Shifted Window Attention. arXiv preprint arXiv:2504.15317.
• Cohen, J. (1968). Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. Psychological bulletin, 70(4), 213.
• Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., ... & Houlsby, N. (2020). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929.
• Fong, D. S., Aiello, L., Gardner, T. W., King, G. L., Blankenship, G., Cavallerano, J. D., Ferris, F. L. ve Klein, R., 2004. Retinopathy in diabetes. Diabetes Care, 27(Suppl 1), S84-S87.
• Elzennary, A., Soliman, M., & Ibrahim, M. (2020, November). Early deep detection for diabetic retinopathy. In 2020 International Symposium on Advanced Electrical and Communication Technologies (ISAECT) (pp. 1-5). IEEE.
• Gonzalez, R. C. (2009). Digital image processing. Pearson education india.
• Gulshan, V., Peng, L., Coram, M., Stumpe, M. C., Wu, D., Narayanaswamy, A., & others. (2016). Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA, 316(22), 2402-2410.
• 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 (pp. 770-778).
• 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 (pp. 4700-4708).
• International Diabetes Federation. (2021). IDF diabetes atlas (10th ed.). Brussels, Belgium: International Diabetes Federation. https://www.diabetesatlas.org
• İncir, R., & Bozkurt, F. (2024). A study on effective data preprocessing and augmentation method in diabetic retinopathy classification using pre-trained deep learning approaches. Multimedia Tools and Applications, 83(4), 12185-12208.
• Kaggle. (2019). APTOS 2019 Blindness Detection [Veri seti]. https://www.kaggle.com/competitions/aptos2019-blindness-detection
• Lavanya, R. V., Sumesh, E. P., Jayakumari, C., & Isaac, R. (2020, November). Detection and classification of diabetic retinopathy using raspberry PI. In 2020 4th International Conference on Electronics, Communication and Aerospace Technology (ICECA) (pp. 1688-1691). IEEE.
• LeCun, Y., Bengio, Y. ve Hinton, G., 2015. Deep learning. Nature, 521(7553), 436-444.
• Lin, T. Y., Goyal, P., Girshick, R., He, K., & Dollár, P. (2017). Focal loss for dense object detection. In Proceedings of the IEEE international conference on computer vision (pp. 2980-2988).
• Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., ... & Guo, B. (2021). Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 10012-10022).
• Majumder, S., & Kehtarnavaz, N. (2021). Multitasking deep learning model for detection of five stages of diabetic retinopathy. IEEE Access, 9, 123220-123230.
• Masters, D., & Luschi, C. (2018). Revisiting small batch training for deep neural networks. arXiv preprint arXiv:1804.07612.
• Minarno, A. E., Bagaskara, A. D., Bimantoro, F., & Suharso, W. (2025). Classification of Diabetic Retinopathy Based on Fundus Image Using InceptionV3. JOIV: International Journal on Informatics Visualization, 9(1), 23-28.
• Nguyen, Q. H., Muthuraman, R., Singh, L., Sen, G., Tran, A. C., Nguyen, B. P., & Chua, M. (2020, January). Diabetic retinopathy detection using deep learning. In Proceedings of the 4th international conference on machine learning and soft computing (pp. 103-107).
• Pamadi, A. M., Ravishankar, A., Nithya, P. A., Jahnavi, G., & Kathavate, S. (2022, March). Diabetic retinopathy detection using mobilenetv2 architecture. In 2022 International Conference on Smart Technologies and Systems for Next Generation Computing (ICSTSN) (pp. 1-5). IEEE.
• Patel, R., & Chaware, A. (2020, June). Transfer learning with fine-tuned MobileNetV2 for diabetic retinopathy. In 2020 international conference for emerging technology (INCET) (pp. 1-4). IEEE.
• Rao, M., Zhu, M., & Wang, T. (2020). Conversion and implementation of state-of-the-art deep learning algorithms for the classification of diabetic retinopathy. arXiv preprint arXiv:2010.11692.
• Saito, T., & Rehmsmeier, M. (2015). The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PloS one, 10(3), e0118432.
• Saranya, P., Devi, S. K., & Bharanidharan, B. (2022, March). Detection of diabetic retinopathy in retinal fundus images using densenet based deep learning model. In 2022 international mobile and embedded technology conference (MECON) (pp. 268-272). IEEE.
• Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
• Sun, H., Saeedi, P., Karuranga, S., Pinkepank, M., Ogurtsova, K., Duncan, B. B., ... & Magliano, D. J. (2022). IDF Diabetes Atlas: Global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes research and clinical practice, 183, 109119.
• Sokolova, M., & Lapalme, G. (2009). A systematic analysis of performance measures for classification tasks. Information Processing & Management, 45(4), 427-437. https://doi.org/10.1016/j.ipm.2009.03.002
• Tan, M., & Le, Q. (2019, May). Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning (pp. 6105-6114). PMLR.
• Taufiqurrahman, S., Handayani, A., Hermanto, B. R., & Mengko, T. L. E. R. (2020, November). Diabetic retinopathy classification using a hybrid and efficient MobileNetV2-SVM model. In 2020 IEEE Region 10 Conference (Tencon) (pp. 235-240). IEEE.
• Thota, N. B., & Reddy, D. U. (2020, August). Improving the accuracy of diabetic retinopathy severity classification with transfer learning. In 2020 IEEE 63rd International Midwest Symposium on Circuits and Systems (MWSCAS) (pp. 1003-1006). IEEE.
• Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. In Advances in Neural Information Processing Systems (Vol. 30). Curran Associates, Inc. https://papers.nips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf
• Yang, J., Shi, R., & Ni, B. (2021, April). Medmnist classification decathlon: A lightweight automl benchmark for medical image analysis. In 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI) (pp. 191-195). IEEE.
• Yang, Y., Cai, Z., Qiu, S., & Xu, P. (2024). Vision transformer with masked autoencoders for referable diabetic retinopathy classification based on large-size retina image. Plos one, 19(3), e0299265.
• Yau, J. W., Rogers, S. L., Kawasaki, R., Lamoureux, E. L., Kowalski, J. W., Bek, T., ... & Meta-Analysis for Eye Disease (META-EYE) Study Group. (2012). Global prevalence and major risk factors of diabetic retinopathy. Diabetes care, 35(3), 556-564.
• Yi, S. L., Yang, X. L., Wang, T. W., She, F. R., Xiong, X., & He, J. F. (2021). Diabetic retinopathy diagnosis based on RA-EfficientNet. Applied Sciences, 11(22), 11035.
• Zuiderveld, K. (1994). Contrast limited adaptive histogram equalization. In Graphics Gems IV (pp. 474–485). Academic Press. https://doi.org/10.1016/B978-0-12-336156-1.50061-6
• Zhang, P., Zhang, X., Brown, J., Vistisen, D., Sicree, R., Shaw, J., & Nichols, G. (2010). Global healthcare expenditure on diabetes for 2010 and 2030. Diabetes research and clinical practice, 87(3), 293-301.
• World Health Organization. (2020). WHO/Europe is focusing on eye screening for people with diabetes. https://www.who.int/europe/news/item/13-11-2020-who-europe-is-focusing-on-eye-screening-for-people-with-diabetes (Erişim Tarihi: 22.11.2024).
• World Health Organization. (2023). Blindness and vision impairment. https://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairmentDünya Sağlık Örgütü (Erişim Tarihi: 22.11.2024).