Vision Transformer and Attention-Augmented Autoencoders with Swarm Intelligence Optimization for Diabetic Retinopathy Detection and Severity Grading: A Systematic Review (۲۰۲۰–۲۰۲۵)

Diabetic retinopathy (DR) remains the leading cause of preventable blindness in working-age adults. Recent advances in vision transformers (ViT), attention-augmented variational and denoising autoencoders, and swarm intelligence-based hyperparameter optimization have dramatically reshaped automated DR detection and five-class severity grading. This systematic review, following PRISMA guidelines, analyzes ۴۰ studies published between ۲۰۲۰ and ۲۰۲۵ that integrate Vision Transformers with generative autoencoder backbones and/or metaheuristic optimization on fundus photography and OCT/OCTA datasets.

کلیدواژه ها:

Diabetic Retinopathy ، Vision Transformer ، Variational Autoencoder