Abstract: Deep neural networks have demonstrated vulnerabilities to black-box adversarial patch attacks in image analysis tasks, raising concerns about their robustness in safety-critical applications ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results