Adversarial Patch

Code

• github.com/Chenglin-Yang/PatchAttack
  pytorch★ 65
• github.com/kenny-co/sgd-uap-torch
  pytorch★ 49
• github.com/eiriniOG/seedtag-codetest
  tf★ 0
• github.com/bingcheng45/hnr-extension
  tf★ 0
• github.com/jhayes14/adversarial-patch
  pytorch★ 0
• github.com/PresentJay/Adversarial_Patch_Attack_With_Pytorch
  pytorch★ 0
• github.com/a-lincui/adversarial_patch_attack
  pytorch★ 0
• github.com/zhaojb17/Adversarial_Patch_Attack
  pytorch★ 0

Abstract

We present a method to create universal, robust, targeted adversarial image patches in the real world. The patches are universal because they can be used to attack any scene, robust because they work under a wide variety of transformations, and targeted because they can cause a classifier to output any target class. These adversarial patches can be printed, added to any scene, photographed, and presented to image classifiers; even when the patches are small, they cause the classifiers to ignore the other items in the scene and report a chosen target class. To reproduce the results from the paper, our code is available at https://github.com/tensorflow/cleverhans/tree/master/examples/adversarial_patch

Reproductions