I'm learning about training images, but when I read the paper of Strebelle
(2002) about SNESIM I got a little confuse, this method is very dependient
of the training image, no matter the hard data, then, if my geological
images has a lot uncertainty my parameter simulation is very uncertainty
too?; What about IMPALA and Enesim.
How should I built a great training image?
I'm trying a issue of connectivity, but never I had worked whit stochastics