Enhancing Retinal Scan Classification: A Comparative Study of Transfer Learning and Ensemble Techniques

Ophthalmic diseases are a significant health concern globally, causing visual impairment and blindness in millions of people, particularly dispersed populations. Among these diseases, retinal fundus leading cause irreversible vision loss, early diagnosis treatment can prevent this outcome. Retinal scans have become an indispensable tool for doctors to diagnose multiple ocular simultaneously. In paper, the results variety deep learning models (DenseNet-201, ResNet125V2, XceptionNet, EfficientNet-B7, MobileNetV2, EfficientNetV2M) ensemble approaches presented, which accurately detect 20 common by analyzing scan images. The proposed model is able achieve remarkable accuracy 96.98% risk classification 76.92% multi-disease detection, demonstrating its potential use clinical settings. By utilizing model, provide swift accurate diagnoses patients, improving their chances receiving timely preserving vision.