Optimized Detection of Diabetic Retinopathy Through Image Preprocessing and Ensemble Models
DOI:
https://doi.org/10.69980/ajpr.v28i4.475Keywords:
Diabetic Retinopathy Detection, Image Preprocessing, Convolutional Neural Networks (CNNs), Ensemble Learning, Data Augmentation, Regularization Techniques.Abstract
One of the most common causes of blindness in diabetics is diabetic retinopathy, thus screening is crucial. Once the issue has been recognized, it's critical to take the appropriate action. This work uses the two-phase experimental scheme to address some of the most difficult problems in DR detection, including image quality, noise, and variability of the DR manifestation.
Phase 1 findings included the adoption of CNN models with poor performance due to overtraining and insufficient preprocessing, as well as basic preprocessing that had issues with generalization. In addition to using various data augmentation techniques and regularization methods like dropout and L2 regularization to eliminate the aforementioned issues, Phase 2 was improved by implementing sophisticated data preprocessing techniques like contrast limited adaptive hue saturation and intensity (CLAHE) to enhance contrast. The CNN models, including MobileNetV2, InceptionV3, and InceptionResNetV2, were further finetuned using ensemble techniques such voting, weighted average, and averaging. Two datasets showed improvements; MobileNetV2 had the best test accuracy, at 96.11, while ensemble approaches enhanced the model's robustness with an AUC of 0.95. To the best of our knowledge, this work provides a better and more efficient combination of advanced preprocessing and ensemble methods for DR detection to support clinical and resource-constrained settings that need early diagnosis and intervention.
References
[1] World Health Organization: WHO & World Health Organization: WHO. (2024, November 14). Diabetes. https://www.who.int/news-room/fact-sheets/detail/diabetes
[2] Karadeniz et.al., International Diabetes Federation, The Fred Hollows Foundation, Bayer Pharma AG, & Novartis Pharma AG. (2015). Diabetes eye health: A guide for health care professionals. International Diabetes Federation. https://idf.org/media/uploads/2023/05/attachments-46.pdf
[3] Sinclair, S. H., & Schwartz, S. S. (2019). Diabetic Retinopathy–An Underdiagnosed and Undertreated Inflammatory, Neuro-Vascular Complication of Diabetes. In Frontiers in Endocrinology (Vol. 10). Frontiers Media SA. https://doi.org/10.3389/fendo.2019.00843
[4] Sun, J.K., Aiello, L.P. (2021). Nonproliferative and Proliferative Diabetic Retinopathy. In: Albert, D., Miller, J., Azar, D., Young, L.H. (eds) Albert and Jakobiec's Principles and Practice of Ophthalmology. Springer, Cham. https://doi.org/10.1007/978-3-319-90495-5_24-1
[5] Kampik, A. (2013). Laser, intravitreal drug application, and surgery to treat diabetic eye disease. In Oman Journal of Ophthalmology (Vol. 6, Issue 4, p. 26). Medknow. https://doi.org/10.4103/0974-620x.122291
[6] Scanlon, P. H. (2017). The English National Screening Programme for diabetic retinopathy 2003–2016. In Acta Diabetologica (Vol. 54, Issue 6, pp. 515–525). Springer Science and Business Media LLC. https://doi.org/10.1007/s00592-017-0974-1
[7] Mansour, R. F. (2017). Deep-learning-based automatic computer-aided diagnosis system for diabetic retinopathy. In Biomedical Engineering Letters (Vol. 8, Issue 1, pp. 41–57). Springer Science and Business Media LLC. https://doi.org/10.1007/s13534-017-0047-y
[8] Panwar, N., Huang, P., Lee, J., Keane, P. A., Chuan, T. S., Richhariya, A., Teoh, S., Lim, T. H., & Agrawal, R. (2016). Fundus Photography in the 21st Century—A Review of Recent Technological Advances and Their Implications for Worldwide Healthcare. In Telemedicine and e-Health (Vol. 22, Issue 3, pp. 198–208). Mary Ann Liebert Inc. https://doi.org/10.1089/tmj.2015.0068
[9] Everett, L. A., & Paulus, Y. M. (2021). Laser Therapy in the Treatment of Diabetic Retinopathy and Diabetic Macular Edema. In Current Diabetes Reports (Vol. 21, Issue 9). Springer Science and Business Media LLC. https://doi.org/10.1007/s11892-021-01403-6
[10] Duffy, A. M., Bouchier-Hayes, D. J., & Harmey, J. H. (2013). Vascular Endothelial Growth Factor (VEGF) and Its Role in Non-Endothelial Cells: Autocrine Signalling by VEGF. Madame Curie Bioscience Database - NCBI Bookshelf. https://www.ncbi.nlm.nih.gov/books/NBK6482/
[11] Dervenis, N., Mikropoulou, A. M., Tranos, P., & Dervenis, P. (2017). Ranibizumab in the Treatment of Diabetic Macular Edema: A Review of the Current Status, Unmet Needs, and Emerging Challenges. In Advances in Therapy (Vol. 34, Issue 6, pp. 1270–1282). Springer Science and Business Media LLC. https://doi.org/10.1007/s12325-017-0548-1
[12] The American Society of Retina Specialists. (n.d.). Vitrectomy - Patients - The American Society of Retina Specialists. ASRS. https://www.asrs.org/patients/retinal-diseases/25/vitrectomy
[13] Mansour, R.F. Deep-learning-based automatic computer-aided diagnosis system for diabetic retinopathy. Biomed. Eng. Lett. 8, 41–57 (2018). https://doi.org/10.1007/s13534-017-0047-y
[14] Doshi, D., Shenoy, A., Sidhpura, D., & Gharpure, P. (2016). Diabetic retinopathy detection using deep convolutional neural networks. In 2016 International Conference on Computing, Analytics and Security Trends (CAST) (pp. 261–266). 2016 International Conference on Computing, Analytics and Security Trends (CAST). IEEE. https://doi.org/10.1109/cast.2016.7914977
[15] Zago, G. T., Andreão, R. V., Dorizzi, B., & Teatini Salles, E. O. (2020). Diabetic retinopathy detection using red lesion localization and convolutional neural networks. In Computers in Biology and Medicine (Vol. 116, p. 103537). Elsevier BV. https://doi.org/10.1016/j.compbiomed.2019.103537
[16] Alyoubi, W. L., Shalash, W. M., & Abulkhair, M. F. (2020). Diabetic retinopathy detection through deep learning techniques: A review. In Informatics in Medicine Unlocked (Vol. 20, p. 100377). Elsevier BV. https://doi.org/10.1016/j.imu.2020.100377
[17] Pradhan, A., Sarma, B., Nath, R. K., Das, A., & Chakraborty, A. (2020). Diabetic Retinopathy Detection on Retinal Fundus Images Using Convolutional Neural Network. In Communications in Computer and Information Science (pp. 254–266). Springer Singapore. https://doi.org/10.1007/978-981-15-6315-7_21
[18] Nguyen, Q. H., Muthuraman, R., Singh, L., Sen, G., Tran, A., Nguyen, B. P., & Chua, M. (2020). Diabetic Retinopathy Detection using Deep Learning. In Proceedings of the 11th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. https://doi.org/10.1145/3380688.3380709
[19] Mishra, S., Hanchate, S. M., & Saquib, Z. (2020). Diabetic Retinopathy Detection using Deep Learning. In 2020 IEEE International Conference on System, Computation, Automation and Networking. https://doi.org/10.1109/ICSTCEE49637.2020.9277506
[20] Gangwar, A., & Ravi, V. (2020). Diabetic Retinopathy Detection Using Transfer Learning and Deep Learning. In Springer Proceedings in Mathematics & Statistics (pp. 799–808). https://doi.org/10.1007/978-981-15-5788-0_64
[21] Tufail, A. B., Ullah, I., Khan, W. U., Asif, M., Ahmad, I., Ma, Y. K., Khan, R., Kalimullah, & Ali, M. S. (2021). Diagnosis of Diabetic Retinopathy through Retinal Fundus Images and 3D Convolutional Neural Networks with Limited Number of Samples. Wireless Communications and Mobile Computing. https://doi.org/10.1155/2021/6013448
[22] Oh, K., Kang, H., Leem, D., Lee, H., Seo, K. Y., & Yoon, S. (2021). Early detection of diabetic retinopathy based on deep learning and ultra-wide-field fundus images. Scientific Reports. https://doi.org/10.1038/s41598-021-81539-3
[23] Dai, L., Wu, L., Li, H., Cai, C., Wu, Q., Kong, H., Liu, R., Wang, X., Hou, X., Liu, Y., Long, X., Wen, Y., Lu, L., Shen, Y., Chen, Y., Shen, D., Yang, X., Zou, H., Sheng, B., & Jia, W. (2021). A deep learning system for detecting diabetic retinopathy across the disease spectrum. Nature Communications. https://doi.org/10.1038/s41467-021-23458-5
[24] Firke, S. N., & Jain, R. B. (2021). Convolutional Neural Network for Diabetic Retinopathy Detection. In 2021 IEEE International Conference on Advances in Information Systems (pp. 425–432). IEEE. https://doi.org/10.1109/ICAIS50930.2021.9395796
[25] Tassanee Hattiya. (2021). Diabetic Retinopathy Detection Using Convolutional Neural Network: A Comparative Study on Different Architectures. Mahasarakham International Journal of Engineering Technology, 7, 50–60. https://doi.org/10.14456/MIJET.2021.8
[26] Das, D., Biswas, S.K. & Bandyopadhyay, S. Detection of Diabetic Retinopathy using Convolutional Neural Networks for Feature Extraction and Classification (DRFEC). Multimed Tools Appl 82, 29943–30001 (2023). https://doi.org/10.1007/s11042-022-14165-4
[27] Ragab, M., AL-Ghamdi, A. S. A.-M., Fakieh, B., Choudhry, H., Mansour, R. F., & Koundal, D. (2022). Prediction of Diabetes through Retinal Images Using Deep Neural Network. In B. Ding (Ed.), Computational Intelligence and Neuroscience (Vol. 2022, pp. 1–6). Hindawi Limited. https://doi.org/10.1155/2022/7887908
[28] Nandakumar, R., Saranya, P., Ponnusamy, V., Hazra, S., & Gupta, A. (2023). Detection of Diabetic Retinopathy from Retinal Images Using DenseNet Models. In Computer Systems Science and Engineering (Vol. 45, Issue 1, pp. 279–292). Tech Science Press. https://doi.org/10.32604/csse.2023.028703
[29] Butt, M. M., Iskandar, D. N. F. A., Abdelhamid, S. E., Latif, G., & Alghazo, R. (2022). Diabetic Retinopathy Detection from Fundus Images of the Eye Using Hybrid Deep Learning Features. In Diagnostics (Vol. 12, Issue 7, p. 1607). MDPI AG. https://doi.org/10.3390/diagnostics12071607
[30] Kusakunniran, W., Karnjanapreechakorn, S., Choopong, P., Siriapisith, T., Tesavibul, N., Phasukkijwatana, N., Prakhunhungsit, S., & Boonsopon, S. (2022). Detecting and staging diabetic retinopathy in retinal images using multi-branch CNN. In Applied Computing and Informatics. Emerald. https://doi.org/10.1108/aci-06-2022-0150
[31] Bhimavarapu, U., & Battineni, G. (2022). Deep Learning for the Detection and Classification of Diabetic Retinopathy with an Improved Activation Function. In Healthcare (Vol. 11, Issue 1, p. 97). MDPI AG. https://doi.org/10.3390/healthcare11010097
[32] Pradhan, A., Sarma, B., Nath, R. K., Das, A., & Chakraborty, A. (2020). Diabetic Retinopathy Detection on Retinal Fundus Images Using Convolutional Neural Network. In Communications in Computer and Information Science (pp. 254–266). Springer Singapore. https://doi.org/10.1007/978-981-15-6315-7_21
[33] Asia, A.-O., Zhu, C.-Z., Althubiti, S. A., Al-Alimi, D., Xiao, Y.-L., Ouyang, P.-B., & Al-Qaness, M. A. A. (2022). Detection of Diabetic Retinopathy in Retinal Fundus Images Using CNN Classification Models. In Electronics (Vol. 11, Issue 17, p. 2740). MDPI AG. https://doi.org/10.3390/electronics11172740
[34] "Nahiduzzaman, Md., Robiul Islam, Md., Omaer Faruq Goni, Md., Shamim Anower, Md., Ahsan, M., Haider, J., & Kowalski, M. (2023). Diabetic retinopathy identification using parallel convolutional neural network based feature extractor and ELM classifier. In Expert Systems with Applications (Vol. 217, p. 119557). Elsevier BV. https://doi.org/10.1016/j.eswa.2023.119557
[35] Malhi, A., Grewal, R., & Pannu, H. S. (2023). Detection and diabetic retinopathy grading using digital retinal images. In International Journal of Intelligent Robotics and Applications (Vol. 7, Issue 2, pp. 426–458). Springer Science and Business Media LLC. https://doi.org/10.1007/s41315-022-00269-5
[36] Kalyani, G., Janakiramaiah, B., Karuna, A., & Prasad, L. V. N. (2021). Diabetic retinopathy detection and classification using capsule networks. In Complex & Intelligent Systems (Vol. 9, Issue 3, pp. 2651–2664). Springer Science and Business Media LLC. https://doi.org/10.1007/s40747-021-00318-9
[37] Chia, M. A., Hersch, F., Sayres, R., Bavishi, P., Tiwari, R., Keane, P. A., & Turner, A. W. (2023). Validation of a deep learning system for the detection of diabetic retinopathy in Indigenous Australians. In British Journal of Ophthalmology (Vol. 108, Issue 2, pp. 268–273). BMJ. https://doi.org/10.1136/bjo-2022-322237
[38] Jain, A., Gupta, R., & Singhal, J. (2024). Diabetic Retinopathy Detection Using Quantum Transfer Learning (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2405.01734
[39] Data Analysis. (2020, January 17). EyePACS. https://www.eyepacs.com/data-analysis
[40] Prasanna Porwal, S. P. (2018). Indian Diabetic Retinopathy Image Dataset (IDRiD) [Dataset]. IEEE Dataport. https://doi.org/10.21227/H25W98
[41] APTOS-2019 dataset. (2021, November 17). Kaggle. https://www.kaggle.com/datasets/mariaherrerot/aptos2019
[42] Decencière et al.Feedback on a publicly distributed database: The Messidor database. Image Analysis & Stereology, v. 33, n. 3, p. 231-234, aug. 2014. ISSN 1854-5165. Available at: http://www.ias-iss.org/ojs/IAS/article/view/1155
[43] DIARETDB1 - Standard Diabetic Retinopathy Database. (2021, January 24). Kaggle. https://www.kaggle.com/datasets/nguyenhung1903/diaretdb1-standard-diabetic-retinopathy-database
[44] S. M. Pizer, R. E. Johnston, J. P. Ericksen, B. C. Yankaskas and K. E. Muller, "Contrast-limited adaptive histogram equalization: speed and effectiveness," [1990] Proceedings of the First Conference on Visualization in Biomedical Computing, Atlanta, GA, USA, 1990, pp. 337-345, doi: 10.1109/VBC.1990.109340
[45] Castillo Benítez, V. E., Castro Matto, I., Mello Román, J. C., Vázquez Noguera, J. L., García-Torres, M., Ayala, J., Pinto-Roa, D. P., Gardel-Sotomayor, P. E., Facon, J., & Grillo, S. A. (2021). Dataset from fundus images for the study of diabetic retinopathy. In Data in Brief (Vol. 36, p. 107068). Elsevier BV. https://doi.org/10.1016/j.dib.2021.107068
[46] Diabetic Retinopathy 224x224 Gaussian Filtered. (2020, February 18). Kaggle. https://www.kaggle.com/datasets/sovitrath/diabetic-retinopathy-224x224-gaussian-filtered
[47] Shorten, C., Khoshgoftaar, T.M. A survey on Image Data Augmentation for Deep Learning. J Big Data 6, 60 (2019). https://doi.org/10.1186/s40537-019-0197-0
[48] Linkon, A. H. Md., Labib, Md. M., Hasan, T., Hossain, M., & Jannat, M.-E.-. (2021). Deep learning in prostate cancer diagnosis and Gleason grading in histopathology images: An extensive study. In Informatics in Medicine Unlocked (Vol. 24, p. 100582). Elsevier BV. https://doi.org/10.1016/j.imu.2021.100582
[49] Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017). Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 31, Issue 1). Association for the Advancement of Artificial Intelligence (AAAI). https://doi.org/10.1609/aaai.v31i1.11231
[50] Simonyan, K., & Zisserman, A. (2014). Very Deep Convolutional Networks for Large-Scale Image Recognition (Version 6). arXiv. https://doi.org/10.48550/ARXIV.1409.1556
[51] Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L.-C. (2018). MobileNetV2: Inverted Residuals and Linear Bottlenecks. arXiv. https://doi.org/10.48550/ARXIV.1801.04381
[52] Empirical Analysis of a Fine-Tuned Deep Convolutional Model in Classifying and Detecting Malaria Parasites from Blood Smears. (2021). In KSII Transactions on Internet and Information Systems (Vol. 15, Issue 1). Korean Society for Internet Information (KSII). https://doi.org/10.3837/tiis.2021.01.009
[53] Nimmisha Shajihan. (2020). Classification of stages of Diabetic Retinopathy using Deep Learning. Unpublished. https://doi.org/10.13140/RG.2.2.10503.62883
[54] Kumar, R., & Indrayan, A. (2011). Receiver operating characteristic (ROC) curve for medical researchers. In Indian Pediatrics (Vol. 48, Issue 4, pp. 277–287). Springer Science and Business Media LLC. https://doi.org/10.1007/s13312-011-0055-4
Downloads
Published
Issue
Section
License
Copyright (c) 2025 American Journal of Psychiatric Rehabilitation
This work is licensed under a Creative Commons Attribution 4.0 International License.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 International License permitting all use, distribution, and reproduction in any medium, provided the work is properly cited.
