ENSEMBLE APPROACH USING DEEP LEARNING AND TRANSFER LEARNING IN CLASSIFICATION OF DIABETIC RETINOPATHY

The proposed deep learning–based diabetic retinopathy detection system provides an efficient and automated approach for retinal image classification. By integrating CLAHE preprocessing, data augmentation, CNN-based feature extraction, ResNet architecture, and ensemble learning, the system aims to achieve accurate and reliable diabetic retinopathy stage detection. The methodology enhances feature learning, improves classification robustness, and reduces prediction variability. Automated retinal screening can support ophthalmologists in early diagnosis, minimize manual workload, and enable large-scale clinical screening. Overall, the proposed framework demonstrates the potential of artificial intelligence and deep learning techniques in improving healthcare accessibility, diagnostic efficiency, and prevention of vision-related complications.

[1]                1. S. Zhu, C. Xiong, Q. Zhong, and Y. Yao, “Diabetic Retinopathy Classification with Deep Learning via Fundus Images: A Short Survey,” IEEE Access, vol. 12, pp. 20540–20558, 2025.   [2]                T.-E.    Tan    and    T.    Y.    Wong,    “Diabetic retinopathy: Looking forward to 2030,” Frontiers in Endocrinology, vol. 14, pp. 1–12, 2024.   [3]                G. U. Nneji, J. Cai, J. Deng, H. N. Monday, M. A. Hossin, and S. Nahar, “Identification of diabetic retinopathy using weighted fusion deep learning    based     on    dual-channel     fundus    scans,” Diagnostics, vol. 13, no. 4, pp. 540–552, 2023.   [4]            T. M. Usman, Y. K. Saheed, D. Ignace, and A. Nsang, “Diabetic retinopathy detection using principal        component analysis         multi-label    feature extraction and classification,” International Journal of Cognitive Computing in Engineering, vol. 4, pp. 78–88, 2023.   [5]          A. M. Oishi, M. Tawfiq-Uz-Zaman, M. B. H. Emon, and S. Momen, “A deep learning approach to diabetic retinopathy classification,” in Cybernetics Perspectives in Systems, Springer, New York, NY, USA, vol. 503, pp. 417–426, 2022.   [6]            H. Sun, P. Saeedi, S. Karuranga, et al., “IDF Diabetes Atlas: Global, regional, and countrylevel diabetes prevalence estimates for 2021 and projections for 2045,” Diabetes Research and Clinical Practice, vol. 183, pp. 109119, 2022.   [7]            Y. Li, L. Zhang, H. Chen, and X. Wang, “Deep convolutional neural network–based automated detection of diabetic retinopathy using retinal fundus images,” Computer Methods and Programs in Biomedicine, vol. 219, pp. 106–118, 2022.   [8]          Bhuiyan, A. Govindaiah, A. Deobhakta, M. Hossain, R. Rosen, and J. Smith, “Automated diabetic retinopathy screening for primary care settings using deep learning,” Intelligent Based Medicine, vol. 5, pp. 100045, 2021.   [9]            R. Sun, Y. Pang, and W. Li, “Efficient image classification using deep convolutional neural networks,” Electronics, vol. 12, no. 4, pp. 1024–1036, 2021.   [10]         J. Krause, V. Gulshan, E. Rahimy, et al., “Grader variability and the importance of reference standards for evaluating machine learning models for diabetic retinopathy,” Ophthalmology, vol. 125, no. 8, pp. 1264–1272, 2020.