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

<table class="table-group" id="tab2"><tr><td><table class="table"><tr><td class="thead-hr" colspan="5"><hr/></td></tr><tr class="thead"><td class="align_left">Method</td><td class="align_center">MSE</td><td class="align_center">PSNR</td><td class="align_center">SSIM</td><td class="align_center">LPIPS</td></tr><tr><td class="thead-hr" colspan="5"><hr/></td></tr><tr><td class="align_left">DAG</td><td class="align_center">970</td><td class="align_center">19.4</td><td class="align_center">0.999</td><td class="align_center">0.311</td></tr><tr><td class="align_left">UEA</td><td class="align_center">114</td><td class="align_center">27.9</td><td class="align_center">0.999</td><td class="align_center">0.353</td></tr><tr><td class="align_left">WPAE</td><td class="align_center">74</td><td class="align_center">30.3</td><td class="align_center">0.999</td><td class="align_center">0.126</td></tr><tr class="table-tr"><td colspan="5"><hr class="tbody-hr"/></td></tr></table></td></tr></table>

<div>Comparison of indicators for each method.</div>

Security and Communication Networks

tab2

Table 2

Table 2: Improving the Imperceptibility of Adversarial Examples Based on Weakly Perceptual Perturbation in Key Regions 