Transfer learning emerges as a powerful methodology within the realm of deep neural networks. This paradigm entails leveraging pre-trained models, which have been saturated on massive datasets, to enhance the training process for new tasks. By migrating the learned representations from the source do