Augment During Training (ADT) pada DenseNet-121: Klasifikasi Tingkat Keparahan Retinopati Diabetik pada Citra Fundus

Willy Cahyadi; Daniel Martomanggolo  Wonohadidjojo

doi:10.29408/edumatic.v9i3.32527

Authors

Willy Cahyadi Program Studi Informatika, Universitas Ciputra Surabaya https://orcid.org/0009-0000-9121-9304
Daniel Martomanggolo Wonohadidjojo Program Studi Informatika, Universitas Ciputra Surabaya https://orcid.org/0000-0003-2259-288X

DOI:

https://doi.org/10.29408/edumatic.v9i3.32527

Keywords:

adaptive augmentation, deep learning, densenet-121, class imbalance, diabetic retinopathy

Abstract

Diabetic Retinopathy (DR) is a severe complication of diabetes mellitus and a leading cause of blindness among the working-age population. Manual diagnosis using fundus images is often subjective and inefficient for mass screening, while automated diagnosis faces challenges due to class imbalance, limiting model performance and generalization. This study aims to develop a DR severity classification model based on the DenseNet-121 architecture by introducing a dynamic augmentation strategy called Augment During Training (ADT). This strategy was evaluated and compared against three other approaches: Baseline, Oversampling Augmentation, and SMOTE. We used the APTOS 2019 Blindness Detection dataset, comprising 3,662 fundus images across five severity levels. The results showed that the model with ADT achieved an accuracy of 83.65% (95% CI: 81.53–85.77%), an F1-score of 83.12%, precision of 83.80%, and recall of 83.65%, outperforming the other three approaches by a margin of 3.27%. Class-wise analysis demonstrated that ADT effectively addressed data imbalance, yielding high performance in the No DR (0.98) and Moderate (0.77) classes, though improvements are still needed for minority classes such as Severe and Proliferative. This study provides a solid computational foundation for developing future early detection tools for Diabetic Retinopathy.

References

Aggarwal, R., Sounderajah, V., Martin, G., Ting, D. S. W., Karthikesalingam, A., King, D., Ashrafian, H., & Darzi, A. (2021). Diagnostic accuracy of deep learning in medical imaging: A systematic review and meta-analysis. Npj Digital Medicine, 4(1), 65. https://doi.org/10.1038/s41746-021-00438-z

Atwany, M. Z., Sahyoun, A. H., & Yaqub, M. (2022). Deep Learning Techniques for Diabetic Retinopathy Classification: A Survey. IEEE Access, 10, 28642–28655. https://doi.org/10.1109/ACCESS.2022.3157632

Benhamza, M., Dahlui, M., & Said, M. A. (2024). Determining direct, indirect healthcare and social costs for diabetic retinopathy management: A systematic review. BMC Ophthalmology, 24(1), 424. https://doi.org/10.1186/s12886-024-03665-6

Gangwar, A. K., & Ravi, V. (2021). Diabetic Retinopathy Detection Using Transfer Learning and Deep Learning. In V. Bhateja, S.-L. Peng, S. C. Satapathy, & Y.-D. Zhang (Eds.), Evolution in Computational Intelligence. (1176), 679–689. Springer Singapore. https://doi.org/10.1007/978-981-15-5788-0_64

Hao, R., Namdar, K., Liu, L., Haider, M. A., & Khalvati, F. (2021). A Comprehensive Study of Data Augmentation Strategies for Prostate Cancer Detection in Diffusion-Weighted MRI Using Convolutional Neural Networks. Journal of Digital Imaging, 34(4), 862–876. https://doi.org/10.1007/s10278-021-00478-7

Hasan, N., Bao, Y., Shawon, A., & Huang, Y. (2021). DenseNet Convolutional Neural Networks Application for Predicting COVID-19 Using CT Image. SN Computer Science, 2(5), 389. https://doi.org/10.1007/s42979-021-00782-7

Hu, A., Schmidt, M. H. H., & Heinig, N. (2024). Microglia in retinal angiogenesis and diabetic retinopathy. Angiogenesis, 27(3), 311–331. https://doi.org/10.1007/s10456-024-09911-1

Khan, M. B., Ahmad, M., Yaakob, S. B., Shahrior, R., Rashid, M. A., & Higa, H. (2023). Automated Diagnosis of Diabetic Retinopathy Using Deep Learning: On the Search of Segmented Retinal Blood Vessel Images for Better Performance. Bioengineering, 10(4), 413. https://doi.org/10.3390/bioengineering10040413

Kumar, S., Asiamah, P., Jolaoso, O., & Esiowu, U. (2025). Enhancing Image Classification with Augmentation: Data Augmentation Techniques for Improved Image Classification (No. arXiv:2502.18691). arXiv. https://doi.org/10.48550/arXiv.2502.18691

Luo, X., Zhang, B., Wang, W., Xu, Y., Lai, Z., Jin, X., & Zhang, D. (2022). A deep convolutional neural network for diabetic retinopathy detection via mining local and long-range dependence. CAAI Transactions on Intelligence Technology, 7(4), 701–713. https://doi.org/10.1049/cit2.12155

Mas’ud, R. A., & Junta, Z. (2024). Optimasi Convolutional Neural Networks untuk Deteksi Kanker Payudara menggunakan Arsitektur DenseNet. Edumatic: Jurnal Pendidikan Informatika, 8(1), 310–318. https://doi.org/10.29408/edumatic.v8i1.25883

Matharaarachchi, S., Domaratzki, M., & Muthukumarana, S. (2024). Enhancing SMOTE for imbalanced data with abnormal minority instances. Machine Learning with Applications, 18, 100597. https://doi.org/10.1016/j.mlwa.2024.100597

Prajnaparamita, I. (2025). Diabetic Retinopathy: A Growing Concern for Diabetes Patients in Indonesia. Journal Scientific of Mandalika (JSM), 6(4), 768–778.

Puteri, V., Lassie, N., & Huda, M. N. (2022). Gambaran Karakteristik Pasien Retinopati Diabetik yang Dilakukan Pembedahan Vitrektomi Di RSKM Padang Eye Center Tahun 2019-2020. Scientific Journal, 1(3), 177–191. https://doi.org/10.56260/sciena.v1i3.40

Santika, E. F. (2024). Indonesia’s Diabetes Prevalence Rises to 11.7% in 2023. Databoks. Retrieved November 17, 2025, from Indonesia’s Diabetes Prevalence Rises to 11.7% in 2023

Shaniaputri, T., Iskandar, E., & Fajriansyah, A. (2022). Prevalensi Retinopati Diabetik di Puskesmas di Bandung Raya Periode Januari 2019-Desember 2020. eJournal Kedokteran Indonesia, 10(1), 39–45. https://doi.org/10.23886/ejki.10.119.39-45

Song, B., Li, S., Sunny, S., Gurushanth, K., Mendonca, P., Mukhia, N., Patrick, S., Gurudath, S., Raghavan, S., Tsusennaro, I., Leivon, S. T., Kolur, T., Shetty, V., Bushan, V., Ramesh, R., Peterson, T., Pillai, V., Wilder-Smith, P., Sigamani, A., … Liang, R. (2021). Classification of imbalanced oral cancer image data from high-risk population. Journal of Biomedical Optics, 26(10). https://doi.org/10.1117/1.JBO.26.10.105001

Ullah, N., Mohmand, M. I., Ullah, K., Gismalla, M. S. M., Ali, L., Khan, S. U., & Ullah, N. (2022). Diabetic retinopathy detection using genetic algorithm-based CNN features and error correction output code SVM framework. Wireless Communications and Mobile Computing, 2022, Article ID 7095528. https://doi.org/10.1155/2022/7095528

Vishwanath, C. S., Divya, K. L., Vamsika, A., Murali, B. S., & Dinesh, B. (2024). Diabetic Retinopathy Classification by Enhanced Image Filtration Techniques Using Deep Learning. 2024 OPJU International Technology Conference (OTCON) on Smart Computing for Innovation and Advancement in Industry 4.0, 1–7. India: IEEE. https://doi.org/10.1109/OTCON60325.2024.10687473

Wahidin, M., Achadi, A., Besral, B., Kosen, S., Nadjib, M., Nurwahyuni, A., Ronoatmodjo, S., Rahajeng, E., Pane, M., & Kusuma, D. (2024). Projection of diabetes morbidity and mortality till 2045 in Indonesia based on risk factors and NCD prevention and control programs. Scientific Reports, 14(1), 5424. https://doi.org/10.1038/s41598-024-54563-2

Yang, Z., Tan, T.-E., Shao, Y., Wong, T. Y., & Li, X. (2022). Classification of diabetic retinopathy: Past, present and future. Frontiers in Endocrinology, 13, 1079217. https://doi.org/10.3389/fendo.2022.1079217

Yi, D., Baltov, P., Hua, Y., Philip, S., & Sharma, P. K. (2024). Compound Scaling Encoder-Decoder (CoSED) Network for Diabetic Retinopathy Related Bio-Marker Detection. IEEE Journal of Biomedical and Health Informatics, 28(4), 1959–1970. https://doi.org/10.1109/JBHI.2023.3313785

Zhao, P. Y., Bommakanti, N., Yu, G., Aaberg, M. T., Patel, T. P., & Paulus, Y. M. (2023). Deep learning for automated detection of neovascular leakage on ultra-widefield fluorescein angiography in diabetic retinopathy. Scientific Reports, 13(1), 9165. https://doi.org/10.1038/s41598-023-36327-6