- The second thing is matching algorithm. Currently I get the feeling
that my gift is fixed on color more than on shape / texture. sometimes I
get perfect results ( e.g. gift finding all appropriate matches I know
of ) but sometimes it doesn't even show an image it most definitely
should be supposed to. How can I tweak my system to try different
methods or make it more sensitive for shape / texture than color ?
I remember having written a document about configuring and hacking
the gift, which is surely too short, but maybe a starting point.
Have you tried the separate normalisation algorithm or which one
did you use when running the GIFT?
I tried both Classical IDF and Separate normalisation but to be honest,
the results did not differ in the least.
Try to fiddle with the parameters to Classical IDF...
:-)
I too have seen that images that I (as a human) sees as pretty much
identical will not appear to be similar to the GIFT. But I'm quite
satisfied with the results anyway.
My results got better and better as I added more images. Using just a
few hundred images, the results were crap. Adding a few thousand made
it promising. Adding almost a hundred thousand images made the results
pretty decent, for my purposes.
Color does seem to be an important part of the features though...
/Jonas
_______________________________________________
help-GIFT mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-gift
