<I just realized my math is slightly weird, but I'm going to suggest
<that any hope that we're going to magically turn a 5MP camera into
<20MP camera is not realistic.
I have easily doubled the number of pixels with good results on occasion but it
is a bit of a stretch to have 3/4 of all pixels made up out of whole cloth as
is the case with neural network based machine learning up res programs. . If
the training set contains super matches for the image, not saying it might not
work. No mathematical algorithm interpolating pixel values is going to give
as nice a result in such a situation. The programs are not really finding
useful resolution hidden in the image like with deconvolution, but totally
creating pixels based on machine learning. A purist might find the former
kosher, but the latter rather objectionable as in rendering a new sky in PS.
Up res, Mike (not as pure as I used to be)
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|