Developing a Python-Based Framework for Eye Disease Classification Using Machine-Deep Learning Techniques for Enhanced Diagnosis

Elderly eye conditions are a serious health issue. As people age, their eye tissues' ability to function normally declines, and the prevalence of ocular disease rises. People with a personal or family history of eye illnesses, those over the age of 60, diabetes, or a history of eye injuries or surgeries are at an even higher risk, making early identification of these conditions all the more crucial. In this work, a DL system for classifying eye diseases is presented utilizing the APTOS2019 Blindness Detection (BD) dataset. In this work proposed DenseNet121 model was also compared with other deep learning and machine learning models, namely: EfficientNet B3, DenseNet121, NB, and CNN, were carried out the research using F1-score, precision, recall, and accuracy measures. The experimental findings show that DenseNet121 outperforms all other models with a 97.68% accuracy, 97% precision, 98% recall, and 97% F1score. A comparative analysis shows the proposed model shows higher performance with a balanced dataset in comparison to other existing models. The results prove DenseNet121 to be viable for DR classification, and the method shows its promise for future medical image analysis as well as aiding the development of diagnostic systems for ophthalmologic diseases.