Skip to content

Chapter 11 #84

@LucBourrat1

Description

@LucBourrat1

Is there a way to save attacked images after having performed adversarial attack?

I tried that:

img_mod = modified_images[0]
Image.fromarray(np.array(img_mod)).save("test.jpg")
img = Image.open("test.jpg")
predict_on_image(torch.tensor(np.array(img)))

And the prediction should be "lemon" but I still get "African elephant".

Then I tried saving as ".png" instead of ".jpg":

img_mod = modified_images[0]
Image.fromarray(np.array(img_mod)).save("test.png")
img = Image.open("test.png")
predict_on_image(torch.tensor(np.array(img)))

And this time I correctly get the label "lemon".

Can you explain why?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions