LaCViT: A Label-Aware Contrastive Fine-Tuning Framework for Vision Transformers | IEEE Conference Publication | IEEE Xplore