Enhancing Facial Beauty Prediction with a Cross-Attention Vision Transformer and Attention-Guided Augmentation | Synapse