EyeFusionNet: A Hybrid CNN-Transformer Network for Automated Diagnosis of Eye Diseases using Colour Fundus Imaging
DOI:
https://doi.org/10.4108/eetismla.10805Keywords:
Deep Learning, Eye fundus, Eye disease clssification, CNN, OpthalmologyAbstract
INTRODUCTION: The timely and accurate diagnosis of eye conditions, such as myopia, cataracts, and diabetic retinopathy, is critical for preventing vision loss. Conventional diagnostic techniques depend on the manual examination of eye fundus images, a process that is often labour-intensive and susceptible to human errors.
OBJECTIVES: The proposed work aims to develop an automated deep learning-based solution for classifying eye diseases using the newly released Eye Disease Image Dataset with 5,335 images and 10 classes, addressing limitations in traditional manual diagnostic techniques.
METHODS: The proposed EyeFusionNet employs a dual-track architecture combining DenseNet169 for detailed local feature extraction and Transformer-iN-Transformer for capturing global context and long-range dependencies. The outputs are fused and refined using efficient channel attention to focus on important regions within the fundus images, with explainable artificial intelligence used to provide visualisations to doctors to validate the diagnosis.
RESULTS: The EyeFusionNet achieved an accuracy of 89% on the publicly available dataset of eye fundus images, outperforming state-of-the-art CNNs and transformer models, showcasing the potential of deep learning for eye fundus image diagnosis.
CONCLUSION: The proposed EyeFusionNet introduces a dual-track architecture for eye disease classification, providing a strong foundation for advancing automated diagnostic tools in clinical settings with the help of deep learning and explainable artificial intelligence to empower clinicians.
Downloads
References
[1] Forrest SL, Mercado CL, Engmann CM, Stacey AW, Hariharan L, Khan S, et al. Does the current Global Health Agenda Lack Vision? Global Health: Science and Practice 2023;11. https://doi.org/10.9745/GHSP-D-22-00091.
[2] Curran K, Peto T, Jonas JB, Friedman D, Kim JE, et al. Global estimates on the number of people blind or visually impaired by diabetic retinopathy: a meta-analysis from 2000 to 2020. Eye 2024;38:2047–57. https://doi.org/10.1038/s41433-024-03101-5.
[3] Gupta P, Fenwick EK, Man REK, Gan ATL, Sabanayagam C, Quek D, et al. Different impact of early and late stages irreversible eye diseases on vision-specific quality of life domains. Sci Rep 2022;12. https://doi.org/10.1038/s41598-022-12425-9.
[4] R.M. B, Vardhan KB, Nidhish M, Kiran C. S, Nahid Shameem D, Sai Charan V. Eye Disease Detection Using Deep Learning Models with Transfer Learning Techniques. ICST Transactions on Scalable Information Systems 2024;11. https://doi.org/10.4108/eetsis.5971.
[5] Kaur I, Ali A. A Complete Study on Machine Learning Algorithms for Medical Data Analysis. Fog Computing for Intelligent Cloud IoT Systems 2024:137–72. https://doi.org/10.1002/9781394175345.ch7.
[6] Suresh NV, Sridhar J, Selvakumar A, Catherine S. Machine Learning Applications in Healthcare. Advances in Healthcare Information Systems and Administration 2024:1–9. https://doi.org/10.4018/979-8-3693-7452-8.ch001.
[7] ALI L, Gun TC, Alhasan W. Comparative Analysis of Machine Learning Algorithms in Enhancing Healthcare Outcomes. EMSJ 2024;8:606–18. https://doi.org/10.59573/emsj.8(3).2024.38.
[8] Akter S, Dwivedi YK, Sajib S, Biswas K, Bandara RJ, Michael K. Algorithmic bias in machine learning-based marketing models. Journal of Business Research 2022;144:201–16. https://doi.org/10.1016/j.jbusres.2022.01.083.
[9] Alzubaidi L, Zhang J, Humaidi AJ, Al-Dujaili A, Duan Y, Al-Shamma O, et al. Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. J Big Data 2021;8. https://doi.org/10.1186/s40537-021-00444-8.
[10] Sarker IH. Machine Learning: Algorithms, Real-World Applications and Research Directions. SN COMPUT SCI 2021;2. https://doi.org/10.1007/s42979-021-00592-x.
[11] Taye MM. Theoretical understanding of Convolutional Neural Network: Concepts, architectures, applications, Future Directions. Computation 2023;11:52. doi:10.3390/computation11030052.
[12] Shamsan A, Senan EM, Shatnawi HS. Automatic classification of colour fundus images for prediction eye disease types based on hybrid features. Diagnostics 2023;13:1706. https://doi.org/10.3390/diagnostics13101706.
[13] Ali G, Dastgir A, Iqbal MW, Anwar M, Faheem M. A hybrid convolutional neural network model for automatic diabetic retinopathy classification from fundus images. IEEE Journal of Translational Engineering in Health and Medicine 2023;11:341–50. https://doi.org/10.1109/jtehm.2023.3282104.
[14] Al-Fahdawi S, Al-Waisy AS, Zeebaree DQ, Qahwaji R, Natiq H, Mohammed MA, et al. Fundus-DeepNet: Multi-label Deep Learning Classification system for enhanced detection of multiple ocular diseases through data fusion of fundus images. Information Fusion 2024;102:102059. https://doi.org/10.1016/j.inffus.2023.102059.
[15] Bernabe O, Acevedo E, Acevedo A, Carreno R, Gomez S. Classification of eye diseases in fundus images. IEEE Access 2021;9:101267–76. https://doi.org/10.1109/access.2021.3094649.
[16] Mahajan P, Uddin S, Hajati F, Moni MA. Ensemble learning for disease prediction: A Review. Healthcare 2023;11:1808. https://doi.org/10.3390/healthcare11121808.
[17] Zhao Z, Zhang K, Hao X, Tian J, Heng Chua MC, Chen L, et al. Bira-net: Bilinear attention net for diabetic retinopathy grading. 2019 IEEE International Conference on Image Processing (ICIP) 2019:1385–9. https://doi.org/10.1109/icip.2019.8803074.
[18] Kim JH, Jo E, Ryu S, Nam S, Song S, Han YS, et al. A deep learning ensemble method to visual acuity measurement using fundus images. Applied Sciences 2022;12:3190. https://doi.org/10.3390/app12063190.
[19] Qummar S, Khan FG, Shah S, Khan A, Shamshirband S, Rehman ZU, et al. A deep learning ensemble approach for diabetic retinopathy detection. IEEE Access 2019;7:150530–9. https://doi.org/10.1109/access.2019.2947484.
[20] Jiang H, Yang K, Gao M, Zhang D, Ma H, Qian W. An interpretable ensemble deep learning model for diabetic retinopathy disease classification. 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) 2019. https://doi.org/10.1109/embc.2019.8857160.
[21] Iman M, Arabnia HR, Rasheed K. A review of deep transfer learning and recent advancements. Technologies 2023;11:40. https://doi.org/10.3390/technologies11020040.
[22] Chen H, Zeng X, Luo Y, Ye W. Detection of diabetic retinopathy using Deep Neural Network. 2018 IEEE 23rd International Conference on Digital Signal Processing (DSP) 2018. https://doi.org/10.1109/icdsp.2018.8631882.
[23] Zannah TB, Abdulla-Hil-Kafi Md, Sheakh MdA, Hasan MdZ, Shuva TF, Bhuiyan T, et al. Bayesian optimized machine learning model for automated eye disease classification from Fundus Images. Computation 2024;12:190. https://doi.org/10.3390/computation12090190.
[24] Wan S, Liang Y, Zhang Y. Deep convolutional neural networks for diabetic retinopathy detection by image classification. Computers & Electrical Engineering 2018;72:274–82. https://doi.org/10.1016/j.compeleceng.2018.07.042.
[25] Aranha GD, Fernandes RA, Morales PH. Deep transfer learning strategy to diagnose eye-related conditions and diseases: An approach based on low-quality fundus images. IEEE Access 2023;11:37403–11. https://doi.org/10.1109/access.2023.3263493.
[26] Rieck C, Mai C, Eisentraut L, Buettner R. A Novel Transformer-CNN Hybrid Deep Learning Architecture for Robust Broad-Coverage Diagnosis of Eye Diseases on Color Fundus Images. IEEE Access 2025;13:156285–300. https://doi.org/10.1109/access.2025.3606334.
[27] Sharmin S, Rashid MR, Khatun T, Hasan MZ, Uddin MS, Marzia. A dataset of color fundus images for the detection and classification of eye diseases. Data in Brief 2024;57:110979. https://doi.org/10.1016/j.dib.2024.110979.
[28] Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely Connected Convolutional Networks. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017. https://doi.org/10.1109/cvpr.2017.243.
[29] Han K, Xiao A, Wu E, Guo J, Xu C, Wang Y. Transformer in Transformer 2021. https://doi.org/10.48550/ARXIV.2103.00112.
[30] Wang Q, Wu B, Zhu P, Li P, Zuo W, Hu Q. ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2020:11531–9. https://doi.org/10.1109/cvpr42600.2020.01155.
[31] Lundberg S, Lee S-I. A Unified Approach to Interpreting Model Predictions 2017. https://doi.org/10.48550/ARXIV.1705.07874.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Akshaj Singh, Sagar Singh Chauhan, Armaano Ajay, Rishi Manda, Prasanna Bharathi
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open access article distributed under the terms of the CC BY-NC-SA 4.0, which permits copying, redistributing, remixing, transformation, and building upon the material in any medium so long as the original work is properly cited.
