Transfer learning stands as a powerful approach within the realm of deep neural networks. This strategy entails leveraging pre-trained models, which have been saturated on massive datasets, to accelerate the training process for new tasks. By transferring the learned representations from the source