Dear all,

I have a curiosity about GANs and what we can see around on article about this argument. In lots of article at example faces image, you are not able to understand if the image is real or is a fake generated by a GAN. My question is, this condition happen always or the fact the output face is impossible to distinguish if true or false to a real face happen only sometimes ? Hope to have been clear in what I am intend and ask.

Thanks,

Fabio