Improving the Imperceptibility of Adversarial Examples Based on Weakly Perceptual Perturbation in Key Regions

<div>(a) The adversarial example generated by UEA and (b) the adversarial example generated by FA. We can see many red noises in adversarial examples and many perturbations in the irrelevant background.</div>

Security and Communication Networks

fig1

Figure 1

Figure 1: Improving the Imperceptibility of Adversarial Examples Based on Weakly Perceptual Perturbation in Key Regions 