Intensified Diabetic Retinopathy Severity Grading: Leveraging Weighted Prominence Learning KNN System

Diabetic retinopathy (DR) represents a significant complication of diabetes that threatens vision due to excess glucose damaging retinal blood vessels. Without timely treatment, this condition can lead to complete vision loss. Early screening and monitoring are crucial for prevention. Several, traditional diagnostic methods employed by ophthalmologists are often labor-intensive, costly, and prone to errors. To enhance DR detection and severity assessment from retinal images, various Artificial Intelligence (AI) methods utilizing computer vision are currently being developed. Despite the promise of AI, many machine learning approaches encounter difficulties with the complex features inherent in retinal images, resulting in low accuracy, particularly in the early stages of DR. In response to these challenges, the proposed work introduces a Weighted Prominence Learning K-Nearest Neighbors (WPL-KNN) method integrated with VGG16 and ResNet50 classifiers, labelled as VGG16-WPL-KNN and ResNet50-WPL-KNN. This new system of assigning weight helps proposed model better capture the intricate aspects, particularly during the initial phases of the illness, resulting in enhanced effectiveness. It utilizes the Indian DR Image Dataset (IDRID), which comprises high-quality retinal images, to improve diabetic retinopathy classification outcomes through effective feature extraction using VGG-16 and ResNet-50.To further enhancing performance, the Seagull Optimization Algorithm (SOA) and Pelican Optimization Algorithm (POA) are applied for hyper parameter tuning. The VGG16-WPL-KNN model achieves an impressive accuracy of 0.99, while the ResNet50-WPL-KNN model achieves an accuracy of 0.96, showcasing the VGG16-WPL-KNN model's accuracy in classifying DR classes. This research contributes to the fields of ophthalmology, pathology, and diabetology by providing a more efficient and accurate means of diagnosing diabetic retinopathy, thereby offering hope for saving the vision of countless people globally.