Skip to content

yangbo93/attack-method

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

50 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RT-DIM

This repository contains code to reproduce results from the paper: Random Transformation of Image Brightness for Adversarial Attack. This paper has been accepted in JOURNAL OF INTELLIGENT & FUZZYSYSTEMS with the according link: https://content.iospress.com/articles/journal-of-intelligent-and-fuzzy-systems/ifs211157.

REQUIREMENTS

  • Environment Anaconda
  • Python 3.7
  • Tensorflow 1.14
  • Numpy 1.18.1
  • cv2 4.4.0.42
  • scipy 1.4.1

Method

we propose a new attack method based on data augmentation that randomly transforms the brightness of the input image at each iteration in the attack process to alleviate overfitting and generate adversarial examples with more transferability. We summarize our algorithm in [Random Transformation of Image Brightness for Adversarial Attack] https://arxiv.org/abs/2101.04321.

Dataset

We use a subset of ImageNet validation set containing 1000 images, most of which are correctly classified by those models.

Models

We use the ensemble of seven models in our submission, many of which are adversarially trained models. The models can be downloaded in (https://github.com/tensorflow/models/tree/master/research/slim, https://github.com/tensorflow/models/tree/archive/research/adv_imagenet_models).

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages