Diabetic Retinopathy Classification Using a Hybrid Deep Learning and Machine Learning Model

Objective: Among diabetic patients, diabetic retinopathy (DR) remains one of the most common causes of preventable blindness and vision loss, making its early detection crucial for preventing irreversible complications. Manual evaluation of fundus photographs is a lengthy process. Additionally, it requires specialized training that is not always available in all clinical settings. Consequently, artificial intelligence‑based automated retinal image analysis systems have emerged as complementary tools to enhance diagnostic accuracy and efficiency. This study proposes an ensemble learning‑based framework to improve the accuracy and robustness of automated DR detection. In the first stage, pretrained convolutional neural network (CNN) models extract high‑level features from fundus images, capturing complex patterns and DR‑related lesions. These features are then fed into several classical machine‑learning classifiers, including Support Vector Machine (SVM), Random Forest, and XGBoost. To further boost discriminative power and reduce classification errors, a stacking ensemble strategy integrates the predictions of the individual classifiers within a meta‑learning framework, enabling the model to learn the optimal combination for DR detection and grading. This hybrid approach effectively combines the strengths of deep learning and classical machine learning, yielding improved performance in DR detection and classification. Experimental results show that the stacking ensemble achieves higher accuracy and F۱‑score compared to individual models, underscoring its potential as an auxiliary tool for early diabetic retinopathy detection.

کلیدواژه ها:

Diabetic retinopathy ، Ensemble learning ، Deep learning ، Machine learning