[Insight-users] Bias-Correcting images

[Torsten Rohlfing]

Arne:

There is a filter in ITK that may do what you want: 
HistogramMatchingImageFilter -- it scales the intensities of one image 
to match the distribution of another. An externally-guided histogram 
equalization if you will. So potentially all you need to do is pick one 
of your images as the reference and run this filter on all other images 
with the selected one as the reference image.

Obviously this procedure will only do what you want it to if the 
relative proportions of tissue volumes are approximately the same in all 
images, so you may run into trouble trying to equalize intensities 
across different subject groups (e.g., younger against older subjects). 
In that case, you may need to do something more sophisticated -- segment 
the images first (e.g., using FAST), determine the individual intensity 
distributions per class and image and then do some sort of fitting to 
get them all to be approximately identical.

Short of that, I am attaching source for a small command line program 
using the existing ITK filter, which may be all you need.

Best,
  Torsten

-- 
Torsten Rohlfing, PhD          SRI International, Neuroscience Program
 Research Scientist             333 Ravenswood Ave, Menlo Park, CA 94025
  Phone: ++1 (650) 859-3379      Fax: ++1 (650) 859-2743
   torsten at synapse.sri.com        http://www.stanford.edu/~rohlfing/

     "Though this be madness, yet there is a method in't"

-------------- next part --------------
A non-text attachment was scrubbed...
Name: IntensityNormalization.cxx
Type: text/x-c++src
Size: 1678 bytes
Desc: not available
Url : http://public.kitware.com/pipermail/insight-users/attachments/20070321/2aeb4231/IntensityNormalization.cxx