"This research tackles the challenge of diversifying face recognition (FR) training via optimized, cost-effective data augmentation (DA). Existing DA techniques lack diversity, prompting the creation of ViT-DiverseGAN, ...