Analysis of abnormalities in retinal blood vessels through deep learning methods is constrained by limited medical data, making augmentation techniques necessary to prevent overfitting. This study proposes the use of Pix2Pix Generative Adversarial Network (GAN) combined with transformation augmentation to increase the number of retinal images and generate synthetic images based on the distribut…