The reasons why Deep Neural Networks are susceptible to being fooled by adversarial examples remains an open discussion. Indeed, many differ- ent strategies can be employed to efficiently generate adversarial attacks, some ...