A system to disrupt the object detection task of deep learning models.
File test_all.py is used to test both state-of-the-art trained models and trained GAN. This is done by passing arguments as required by file. One can perform black-box and white-box attacks trying on different combination of dicriminator and generator.
File attacks.py contains definition of classes for FGSM and DCGAN.
Inorder to train GAN use example.py file.