Test on a single Already Perturbed Image

kabkabm / defensegan

Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models (published in ICLR2018)

Apache License 2.0

229 stars 62 forks source link

Test on a single Already Perturbed Image #5

Open imransalam opened 5 years ago

imransalam commented 5 years ago

Is there any way to use the already trained celeba model (trained with train.py) and infer for a single perturbed image and see if the model regenerates the original?

bibin-sebastian commented 5 years ago

are you looking for reconstructing one image ?

imransalam commented 5 years ago

@bibin-sebastian Yes, I have already perturbed the image with FGSM, I would want to normalize it with a celeba trained network (trained with train.py).