T2T-ViT: A Novel Semantic Image Mining Approach for Improving CBIR Using Vision Transformer | Synapse