"Prediction: Positive ==> True"

Question

"Prediction: Positive ==> True"

Closed this issue 5 years ago · 4 comments

Hi,
i run the demo/compare-prediction.ipynb and i have no idea about

like "Prediction: Positive ==> True" it means the pic is authentic?
and "Prediction: Positive ==> False" it means the pic is tampering?

Answer 1 · 2019-03-05T09:23:42.000Z

Prediction: Positive ==> True" means: The model predicts image as Positive (the image is tampered) and the ground truth is also Positive. So the prediction is True.

Prediction: Positive ==> False" means: The model predicts image as Positive (the image is tampered) but the ground truth is Negative. So the prediction is False.

Answer 2 · 2019-03-05T09:33:03.000Z

Prediction: Positive ==> True" means: The model predicts image as Positive (the image is tampered) and the ground truth is also Positive. So the prediction is True.

Prediction: Positive ==> False" means: The model predicts image as Positive (the image is tampered) but the ground truth is Negative. So the prediction is False.

that means just use decision == 1 can know the image is tampered or not?

Answer 3 · 2019-03-05T13:41:43.000Z

(decision==1) ===> Positive ===> Tampered image
(decision==0) ===> Negative ===> Authentic image

Answer 4 · 2019-03-06T01:32:41.000Z

(decision==1) ===> Positive ===> Tampered image
(decision==0) ===> Negative ===> Authentic image

Thank you