Swarm-Optimized Deep Convolutional Networks for Precise Detection and Grading of Diabetic Retinopathy

Diabetic Retinopathy (DR) is a very critical problem of diabetes that can lead to blindness if not detected and treated in the early stage. This research was conducted on this difficult disease based on image processing and Machine intelligence. In this research, our model detects and classifies the DR using the swarm-optimized deep convolutional networks. The proposed model integrated the convolutional neural networks (CNNs) along with Particle Swarm Optimization (PSO) to enrich the performance of the model execution and results with fine-tuned hyperparameters. This custom-optimized CNN model, designed for image classification, was evaluated on the Kaggle diabetic retinopathy dataset. The swarm-optimized deep CNN model presented balanced performance across all DR classes and achieved a precision of 0.93, recall of 0.97, and F1-score of 0.95 for all classes. The model's overall accuracy reached 0.84 on the selected dataset.