Early detection of ophthalmic diseases is essential for preventing irreversible vision loss and improving patients’ quality of life. Recent advances in artificial intelligence (AI) and medical imaging have created new opportunities for automated screening and clinical decision support. This study proposes an AI-based model for early diagnosis and prediction of ophthalmic diseases using digital retinal images. The proposed approach employs convolutional neural networks to automatically extract visual features from fundus photographs and generate probabilistic predictions of pathological conditions. Digital retinal images were preprocessed through normalization, noise reduction, and contrast enhancement to improve feature representation. The model was trained and validated using separated training and testing subsets. Performance was evaluated using standard metrics including accuracy, recall, F1-score, and receiver operating characteristic (ROC) analysis. Experimental results demonstrate that the proposed model is capable of distinguishing healthy and pathological retinal images and estimating disease progression tendencies. The findings suggest that AI-based retinal image analysis can serve as a supportive tool for early diagnosis and prognosis, potentially reducing diagnostic subjectivity and improving screening efficiency. The proposed system is particularly promising for telemedicine and remote screening applications.
Yakubov et al. (Sat,) studied this question.