Open AtefehBaniasadi opened 3 years ago
Hi
Thank you for your rapid response.
How can I use those scores to get the final decision? Indeed, I would like to know what is the label of each image (Real or Fake)?
Which one of these values (M, s, b, C, T ) can help me know if the image is live or spoof
Many thanks,
you need to use those M,s,b,C,T values in the formula i have attached above. Substitute M in equation 3 and use s,b,C,T in g(I) equation.
Find the score by substituting the values in score equation
Is there a python script in this repo to get this value(M,s,b,C,T) and compute the score?
nope. it is not possible, you need to set alpha0 value by testing empirically, based on the dataset. you can use this code as a starter to train the dataset and test.
Calculating the final score seems a little complicated.
How do we set alpha empirically? any thumb rule?
Hi @santo4ul did you get clue to calculate score ?
@Kapil-23 did you get clue to calculate score ?
Hi
First of all, thank you for sharing your codes
I ran your code with a pre-trained model on my own dataset and for one of the images I got this score "0.270, -0.016, -0.008, -0.015, -0.006", Could you please what these mean?
Thanks,