Deep neural networks (DNNs) are susceptible to adversarial attacks, including the recently introduced locally visible adversarial patch attack, which achieves a success rate exceeding 96%. These attacks pose significant challenges to DNN security. Va...