Feature-Level Fusion of DenseNet121 and EfficientNetV2 with XGBoost for Multi-Class Retinal Classification
DOI:
10.33395/sinkron.v10i1.15670Keywords:
DenseNet121, EfficientNetV2, Feature-Level Fusion, Retinal Fundus Classification, XGBoostAbstract
Accurate and efficient classification of retinal fundus images plays a critical role in supporting the early diagnosis of ocular diseases. However, models relying on a single deep learning backbone often struggle to capture the multi-scale and heterogeneous characteristics of retinal lesions, leading to unstable performance across visually similar disease classes. To address this limitation, this study proposes a novelty feature-level fusion framework that integrates complementary representations from DenseNet121 and EfficientNetV2-s, followed by classification using XGBoost. The fusion pipeline extracts 1024-dimensional features from DenseNet121 and 1280-dimensional features from EfficientNetV2-s, which are concatenated into a unified 2304-dimensional feature vector. Experiments were conducted on a dataset of 10,247 retinal fundus images spanning six categories: Central Serous Chorioretinopathy, Diabetic Retinopathy, Macular Scar, Retinitis Pigmentosa, Retinal Detachment, and Healthy. The proposed fusion model achieved an accuracy of 91.60%, outperforming DenseNet121 XGBoost (91.31%) and EfficientNetV2-s XGBoost (89.70%). Moreover, the fusion strategy demonstrated improved class-level stability, particularly for visually similar retinal disorders where single-backbone models exhibited higher misclassification rates. This study contributes a lightweight yet effective multi-backbone feature-level fusion approach that enhances discriminative representation and classification stability without increasing model complexity. In addition, the use of XGBoost introduces a tree-based decision mechanism that is inherently more interpretable than conventional fully connected layers, offering potential advantages for clinical analysis. Overall, the results highlight the effectiveness of multi-backbone feature fusion as a reliable strategy for automated retinal disease classification.
Downloads
References
Abdullah, A. A., Aldhahab, A., & Abboodi, H. M. Al. (2024). Deep-Ensemble Learning Models for the Detection and Classification of Eye Diseases Based on Engineering Feature Extraction with Efficientb6 and Densnet169. International Journal of Intelligent Engineering and Systems, 17(6), 1001–1022. https://doi.org/10.22266/ijies2024.1231.75
Al-antary, M. T., & Arafa, Y. (2021). Multi-Scale Attention Network for Diabetic Retinopathy Classification. IEEE Access, 9(March), 54190–54200. https://doi.org/10.1109/ACCESS.2021.3070685
Al-fahdawi, S., Al-waisy, A. S., Qader, D., Qahwaji, R., Natiq, H., Abed, M., Nedoma, J., Martinek, R., & Deveci, M. (2024). Fundus-DeepNet : Multi-label deep learning classification system for enhanced detection of multiple ocular diseases through data fusion of fundus images. Information Fusion, 102(September 2023), 102059. https://doi.org/10.1016/j.inffus.2023.102059
Amir, A., Jan, T., Zafar, M. H., & Khattak, S. K. (2025). Sophisticated Ensemble Deep Learning Approaches for Multilabel Retinal Disease Classification in Medical Imaging. CAAI Transactions on Intelligence Technology, 10(4), 1159–1173. https://doi.org/10.1049/cit2.70012
Aslam, A., Farhan, S., Khaliq, M. A., Anjum, F., Afzaal, A., & Kanwal, F. (2023). Convolutional Neural Network-Based Classification of Multiple Retinal Diseases Using Fundus Images. Intelligent Automation and Soft Computing, 36(3), 2607–2622. https://doi.org/10.32604/iasc.2023.034041
Bernabe, O., Acevedo, E., Acevedo, A., Carreno, R., & Gomez, S. (2021). Classification of Eye Diseases in Fundus Images. IEEE Access, 9(April), 101267–101276. https://doi.org/10.1109/ACCESS.2021.3094649
Ho, E., Wang, E., Youn, S., Sivajohan, A., Lane, K., Chun, J., & Hutnik, C. M. L. (2022). Deep Ensemble Learning for Retinal Image Classification. Translational Vision Science and Technology, 11(10), 1–11. https://doi.org/https://doi.org/10.1167/tvst.11.10.39
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017). Densely connected convolutional networks. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2261–2269. https://doi.org/10.1109/CVPR.2017.243
Laurik-feuerstein, L., Sapahia, R., & Debuc, D. C. (2022). The assessment of fundus image quality labeling reliability among graders with different backgrounds. PLOS ONE, 7(17), 1–11. https://doi.org/10.1371/journal.pone.0271156
Li, Z., Xu, M., Yang, X., & Han, Y. (2022). Multi-Label Fundus Image Classification Using Attention Mechanisms and Feature Fusion. Micromachines, 13(6). https://doi.org/10.3390/mi13060947
Liu, S., Wang, W., Deng, L., & Xu, H. (2024). Cnn-trans model: A parallel dual-branch network for fundus image classification. Biomedical Signal Processing and Control, 96(PB), 106621. https://doi.org/10.1016/j.bspc.2024.106621
Lu, Z., Miao, J., Dong, J., Zhu, S., Wu, P., Wang, X., & Feng, J. (2023). Automatic Multilabel Classification of Multiple Fundus Diseases Based on Convolutional Neural Network With Squeeze-and-Excitation Attention. Translational Vision Science and Technology, 12(1), 1–13. https://doi.org/10.1167/tvst.12.1.22
Mustafa, H., Ali, S. F., Bilal, M., & Hanif, M. S. (2022). Multi-Stream Deep Neural Network for Diabetic Retinopathy Severity Classification under a Boosting Framework. IEEE Access, 10(October), 113172–113183. https://doi.org/10.1109/ACCESS.2022.3217216
Naveen, K. V., Anoop, B. N., Siju, K. S., Kar, M. K., & Venugopal, V. (2025). EffNet-SVM: A Hybrid Model for Diabetic Retinopathy Classification Using Retinal Fundus Images. IEEE Access, 13(April), 79793–79804. https://doi.org/10.1109/ACCESS.2025.3566073
Nazih, W., Aseeri, A. O., Atallah, O. Y., & El-Sappagh, S. (2023). Vision Transformer Model for Predicting the Severity of Diabetic Retinopathy in Fundus Photography-Based Retina Images. IEEE Access, 11(October), 117546–117561. https://doi.org/10.1109/ACCESS.2023.3326528
Ouda, O., Abdelmaksoud, E., El-aziz, A. A. A., & Elmogy, M. (2022). Multiple Ocular Disease Diagnosis Using Fundus Images Based on Multi-Label Deep Learning Classification. Electronics, 11(13), 1–27. https://doi.org/10.3390/electronics11131966
Qi, T., Liu, H., Fruhn, L., Low, K., Cursiefen, C., & Prokosch, V. (2025). Understanding Glaucoma : Why it Remains a Leading Cause of Blindness Worldwide. Klinische Monatsblätter Für Augenheilkunde, 07(242), 712–717. https://doi.org/10.1055/a-2617-1575
Rashid, M. R., Sharmin, S., Khatun, T., Hasan, M. Z., & Uddin, M. S. (2024). Eye Disease Image Dataset. Mendeley Data. https://doi.org/10.17632/s9bfhswzjb.1
Rodriguez, M. A., Almarzouqi, H., & Liatsis, P. (2023). Multi-Label Retinal Disease Classification Using Transformers. IEEE Journal of Biomedical and Health Informatics, 27(6), 2739–2750. https://doi.org/10.1109/JBHI.2022.3214086
Sampath, M., Khan, M. A., & Scholar, R. (2025). Efficientvit: A Hybrid CNN-Transformer Framework With Cross-Attention Fusion For Clinically Interpretable Diabetic Retinopathy Grading. International Journal of Creative Research Thoughts (IJCRT), 13(April), 2320–2882. https://doi.org/10.1729/Journal.44896
Tan, M., & Le, Q. V. (2021). EfficientNetV2: Smaller Models and Faster Training. Proceedings of Machine Learning Research, 139, 10096–10106. https://doi.org/10.48550/arXiv.2104.00298
Zhang, P., Jia, Y., & Shang, Y. (2022). Research and application of XGBoost in imbalanced data. International Journal of Distributed Sensor Networks, 18(6). https://doi.org/10.1177/15501329221106935
Downloads
|
How to Cite
Issue
Section
License
Copyright (c) 2026 Jovansa Putra Laksana, Yohannes
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Google Scholar 
Moraref
PKP Index
Garuda Kemdikbud
Indonesia OneSearch
OCLC Worldcat
Dimensions
Index Copernicus
Scilit
Muhammad Khoiruddin harahap
