Hybrid-based predictive model for early detection of Myopia

This study presents a hybrid-based predictive model for early detection of Myopia to enhance ophthalmic diagnostics. The proposed system was developed using a hybrid neural network in order to improve the early identification of myopia condition in patients. The model was trained and tested using a concept that combines several CNN learners that improved model prediction accuracy. The introduction of penalty terms and a user notification mechanism improved the model's ability to deal with complexity problems. A penalty term was introduced in order to make the model converge more quickly with better accuracy because it gives the user control over the layer's output. The hybrid framework was discouraged from utilizing larger weights by adding a penalty term that was based on the network weights' values. The hybrid CNN input and output layers were invariably fitted with a penalty term. The existing single CNN model achieved an accuracy of 84.89%, while the hybrid model outperformed it with a 95.91% detection accuracy. The current CNN had the lowest detection accuracy, and the system was never made better by adding more training examples. These results demonstrate the effectiveness of the proposed approach in improving early detection of Myopia, offering a scalable and accurate solution for medical diagnosis and intervention.