I was trying to do the inhomogeneity correction of T1-weighted MRI brain
data sets. I remove the bones and outer material according to the manual
labeled data. Then I use FAST to do the inhomogeneity correction. I ploted
the histogram before and after the inhomogeneity correction. I found that
in some cases, the ratio of the GM/WM changed much. Before the correction
the ratio of GM/WM is greater than one, but after the correction the ratio
is less than one. The average ground truth ratio is around 1.2-1.3.
Another observation is that the white matter peaks shifted much.
Therefore, I am confused and wondering if anyone else met the same problem
before? Is there any parameters to adjust the ratio and peak positions?
Thanks a lot!
|