Ensemble Model for ClassifyingDiabetic Retinopathy Severity on Fundus Image: Model Ensemble untuk Klasifikasi Tingkat Keparahan Retinopati Diabetik pada Citra Fundus

Prayer Yosua Pua; Oktavian Lantang; Agustinus Jacobus

doi:10.35793/k70bdt82

Authors

Prayer Yosua Pua Universitas Sam Ratulangi Author
Oktavian Lantang Universitas Sam Ratulangi Author
Agustinus Jacobus Universitas Sam Ratulangi Author

DOI:

https://doi.org/10.35793/k70bdt82

Keywords:

Diabetic Retinopathy, Ensemble Model, CNN, Fundus Image, Deep Learning

Abstract

Abstract — Diabetic retinopathy (DR) is a complication of diabetes that is the leading cause of blindness in the world. This research develops an automatic classification system of DR severity using an ensemble model of three CNN architectures: InceptionV3, ResNet152, and DenseNet121. The ensemble voting and stacking methods are applied to improve classification accuracy based on eye fundus images. The dataset used contains 34,152 fundus images with five levels of RD severity. Experimental results show that the DenseNet121 individual model achieved the highest validation accuracy of 95.69%, while ensemble voting improved the performance to 96.45% and ensemble stacking reached 97.26%. The system has been implemented in a Flask-based web application to support early diagnosis of DR in health facilities with limited eye specialists.

Key words— Diabetic Retinopathy, Ensemble Model, CNN, Fundus Image, Deep Learning

Abstrak — Retinopati diabetik (RD) merupakan komplikasi diabetes yang menjadi penyebab utama kebutaan di dunia. Penelitian ini mengembangkan sistem klasifikasi otomatis tingkat keparahan RD menggunakan ensemble model dari tiga arsitektur CNN: InceptionV3, ResNet152, dan DenseNet121. Metode ensemble voting dan stacking diterapkan untuk meningkatkan akurasi klasifikasi berdasarkan citra fundus mata. Dataset yang digunakan berisi 34.152 gambar fundus dengan lima tingkat keparahan RD. Hasil eksperimen menunjukkan bahwa model individual DenseNet121 mencapai akurasi validasi tertinggi sebesar 95,69%, sementara ensemble voting meningkatkan performa menjadi 96,45% dan ensemble stacking mencapai 97,26%. Sistem telah diimplementasikan dalam aplikasi web berbasis Flask untuk mendukung diagnosis dini RD di fasilitas kesehatan dengan keterbatasan spesialis mata.

Kata kunci — Retinopati Diabetik, Ensemble Model, CNN, Citra Fundus, Deep Learning

References

[1] I. Noventi and S. Damawiyah, “FAKTOR RESIKO RETINOPATI DIABETIKA : A CASE – CONTROL,” Indones. J. Health Sci., vol. 10, no. 2, p. 1, Dec. 2018, doi: 10.32528/ijhs.v10i2.1851.

[2] A. Sebastian, O. Elharrouss, S. Al-Maadeed, and N. Almaadeed, “A Survey on Deep-Learning-Based Diabetic Retinopathy Classification,” Diagnostics, vol. 13, no. 3, p. 345, Jan. 2023, doi: 10.3390/diagnostics13030345.

[3] G. Litjens et al., “A survey on deep learning in medical image analysis,” Med. Image Anal., vol. 42, pp. 60–88, Dec. 2017, doi: 10.1016/j.media.2017.07.005.

[4] D. Juniati and A. E. Suwanda, “KLASIFIKASI PENYAKIT MATA BERDASARKAN CITRA FUNDUS RETINA MENGGUNAKAN DIMENSI FRAKTAL BOX COUNTING DAN FUZZY K-MEANS,” Prox. J. Penelit. Mat. Dan Pendidik. Mat., vol. 5, no. 1, pp. 10–18, Feb. 2022, doi: 10.30605/proximal.v5i1.1623.

[5] A. Mujumdar and V. Vaidehi, “Diabetes Prediction using Machine Learning Algorithms,” Procedia Comput. Sci., vol. 165, pp. 292–299, 2019, doi: 10.1016/j.procs.2020.01.047.

[6] D. Sisodia and D. S. Sisodia, “Prediction of Diabetes using Classification Algorithms,” Procedia Comput. Sci., vol. 132, pp. 1578–1585, 2018, doi: 10.1016/j.procs.2018.05.122.

[7] K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA: IEEE, June 2016, pp. 770–778. doi: 10.1109/CVPR.2016.90.

[8] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, “Densely Connected Convolutional Networks,” in 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI: IEEE, July 2017, pp. 2261–2269. doi: 10.1109/CVPR.2017.243.

[9] U. V. Shukla and K. Tripathy, “Diabetic Retinopathy,” in StatPearls, Treasure Island (FL): StatPearls Publishing, 2025. Accessed: Aug. 09, 2025. [Online]. Available: http://www.ncbi.nlm.nih.gov/books/NBK560805/

[10] R. Indraswari, W. Herulambang, and R. Rokhana, “Deteksi Penyakit Mata Pada Citra Fundus Menggunakan Convolutional Neural Network (CNN),” Techno.Com, vol. 21, no. 2, pp. 378–389, May 2022, doi: 10.33633/tc.v21i2.6162.

[11] V. Gulshan et al., “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs,” JAMA, vol. 316, no. 22, p. 2402, Dec. 2016, doi: 10.1001/jama.2016.17216.

[12] O. Russakovsky et al., “ImageNet Large Scale Visual Recognition Challenge,” Int. J. Comput. Vis., vol. 115, no. 3, pp. 211–252, Dec. 2015, doi: 10.1007/s11263-015-0816-y.

[13] G. Quellec, M. Lamard, P. M. Josselin, G. Cazuguel, B. Cochener, and C. Roux, “Optimal Wavelet Transform for the Detection of Microaneurysms in Retina Photographs,” IEEE Trans. Med. Imaging, vol. 27, no. 9, pp. 1230–1241, Sept. 2008, doi: 10.1109/TMI.2008.920619.

[14] J. Wang, S. Mall, and L. Perez, “The Effectiveness of Data Augmentation in Image Classiﬁcation using Deep Learning”.

[15] C. Szegedy et al., “Going Deeper with Convolutions,” 2014, arXiv. doi: 10.48550/ARXIV.1409.4842.

[16] A. Ahmad, “Mengenal Artiﬁcial Intelligence, Machine Learning, Neural Network, dan Deep Learning,” 2017.

[17] M. Aryuni, “Penerapan Ensemble Feature Selection dan Klasterisasi Fitur pada Klasifikasi Dokumen Teks,” ComTech Comput. Math. Eng. Appl., vol. 4, no. 1, p. 333, June 2013, doi: 10.21512/comtech.v4i1.2745.

[18] S. E. Limantoro, Y. Kristian, and D. D. Purwanto, “Pemanfaatan Deep Learning pada Video Dash Cam untuk Deteksi Pengendara Sepeda Motor,” J. Nas. Tek. Elektro Dan Teknol. Inf. JNTETI, vol. 7, no. 2, June 2018, doi: 10.22146/jnteti.v7i2.419.

[19] T. R. Mahesh, V. Vinoth Kumar, V. Vivek, K. M. Karthick Raghunath, and G. Sindhu Madhuri, “Early predictive model for breast cancer classification using blended ensemble learning,” Int. J. Syst. Assur. Eng. Manag., vol. 15, no. 1, pp. 188–197, Jan. 2024, doi: 10.1007/s13198-022-01696-0.

[20] R. Egele, I. Guyon, Y. Sun, and P. Balaprakash, “Is One Epoch All You Need For Multi-Fidelity Hyperparameter Optimization?,” Sept. 26, 2023, arXiv: arXiv:2307.15422. doi: 10.48550/arXiv.2307.15422.

[21] O. A. Lantang, R. E. C. Sendow, and F. D. Kambey, “Implementation of Bidirectional Long Short-Term Memory and Convolutional Neural Network in Detecting Hoax Content on Social Media,” Appl. Inf. Syst. Manag. AISM, vol. 8, no. 1, pp. 151–158, May 2025, doi: 10.15408/aism.v8i1.45222.

[22] L. M. Fleuren et al., “Machine learning for the prediction of sepsis: a systematic review and meta-analysis of diagnostic test accuracy,” Intensive Care Med., vol. 46, no. 3, pp. 383–400, Mar. 2020, doi: 10.1007/s00134-019-05872-y.

[23] S. Clara, D. L. Prianto, R. A. Habsi, E. F. Lumbantobing, S. Kom, and M. Kom, “Implementasi Seleksi Fitur Pada Algoritma Klasifikasi Machine Learning Untuk Prediksi Penghasilan Pada Adult Income Dataset,” 2021.

[24] W. Abdun Naseer, S. Sarwido, and B. B. Wahono, “GRADIENT BOOSTING OPTIMIZATION WITH PRUNING TECHNIQUE FOR PREDICTION OF BMT AL-HIKMAH PERMATA CUSTOMER DATA,” J. Inform. Teknol. Dan Sains Jinteks, vol. 6, no. 3, pp. 719–727, Aug. 2024, doi: 10.51401/jinteks.v6i3.4702.

[25] A. Nurizki, A. Fitrianto, and A. Mohamad Soleh, “Performance of Ensemble Learning in Diabetic Retinopathy Disease Classification,” Sci. J. Inform., vol. 11, no. 2, pp. 375–386, May 2024, doi: 10.15294/sji.v11i2.4725.

[26] P. Mahajan, S. Uddin, F. Hajati, and M. A. Moni, “Ensemble Learning for Disease Prediction: A Review,” Healthcare, vol. 11, no. 12, p. 1808, June 2023, doi: 10.3390/healthcare11121808.

[27] F. H. Syahrul and P. S. Sasongko, “Penerapan Convolutional Neural Network Untuk Klasifikasi Tingkat Keparahan Retinopati Diabetik Pada Penderita Diabetes Melitus,” J. Masy. Inform., vol. 13, no. 1, pp. 1–14, May 2022, doi: 10.14710/jmasif.13.1.42354.

[28] R. Faurina, E. P. Purwandari, M. T. Pratama, and I. Agustian, “Klasifikasi Level Non-Proliferatif Retinopati Diabetik Dengan Ensemble Convolutional Neural Network,” Pseudocode, vol. 8, no. 1, pp. 1–10, Mar. 2021, doi: 10.33369/pseudocode.8.1.1-10.

[29] K. Hendrawan, A. T. Handayani, A. Andayani, E. Titiek, and A. B. Gumelar, “Classification of diabetic retinopathy using ensemble convolutional neural network architectures,” Universa Med., vol. 43, no. 2, pp. 188–194, July 2024, doi: 10.18051/UnivMed.2024.v43.188-194.