Attention-Based Knowledge Distillation in Scene Recognition: The Impact of a DCT-Driven Loss | Publicación