So, what do you think?
 

"Rich A" wrote

| http://webdav.tuebingen.mpg.de/pixel/enhancenet/

I don't understand why others think this is silly.
Did I miss something? Is it not promising that an
algorythm could offer a sharpness not available
via normal resizing methods? This seems to be a
kind of "smart contrast" method.

I'd be interested to find source code to possibly
add such a method to resizing operations. (I have
neither the patience nor the expertise to deduce it
from thir published explanations, but their comparison
to bicubic [arguably the best enlarging algorythm]
looks promising.)