Prev: imregionalmax
Next: combine image parts
From: ImageAnalyst on 4 Sep 2009 15:09
H:
Sounded pretty easy until your last sentence. Just find the pixels >
0 and then find the max and min of those, and assign them in a linear
way, using intlut(), to some desired output range.

However if you're talking about matching not just linearly between the
endpoints, but also taking about gamma then you're really talking
about a complete histogram match. You can use histeq() as a start.
It's easy because it's built in to the Image Processing Toolbox, but I
believe it just uses the standard method (CDF transform)which is
actually not very good. You can try sorting and sampling, or adding a
half graylevel of noise before transforming by the CDF - two other
methods that give better results. I don't really want to get into the
details of histogram matching. You can check the internet for
histogram matching, or look over in sci.image.processing where it's
been discussed before.
Good luck,
ImageAnalyst
 | 
Pages: 1
Prev: imregionalmax
Next: combine image parts