Optimizing vision care: Dual path network model in eye disease classification

The increasing prevalence and severity of eye diseases worldwide underscore the urgent need for advanced diagnostic tools and interventions to address the growing burden on global public health. The study on eye disease classification holds significant relevance due to its potential impact on enhancing early detection, diagnosis, and treatment of various ocular conditions. Timely and accurate identification of eye diseases such as cataracts, glaucoma and diabetic retinopathy is crucial for preventing vision loss and improving overall patient outcomes. This innovative model leverages the synergistic features of the Dual Path Network (DPN) due to its ability to capture both global and local features simultaneously, enabling more comprehensive information extraction and contributing to enhanced model performance. The dataset, sourced from Kaggle, comprises retinal images categorized into Normal, Diabetic Retinopathy, Cataract, and Glaucoma, each class containing approximately 1000 images. The proposed model exhibits robust performance across various metrics, achieving a commendable accuracy of 95.32 %, showcasing its proficiency in accurate classification. The notable capacity of the model to reduce false positives while effectively identifying positive cases underscores its potential as a valuable tool for the early detection of eye diseases. By employing advanced technology like a dual-path network, the study aims to provide a more sophisticated and efficient tool for automated classification, thereby assisting healthcare professionals in making informed decisions.