How to normalize the healthy/MS image pair

[naga-karthik]

A major issue during lesion copy-pasting is the differnence in the intensities of the healthly and MS images. Due to this difference, what's hypo/hyperintense in the MS image is not necessarily the same in the healthy subject.

I have tried using (Global) normalization methods, namely, min-max (or, 0-1) and z-score normalization. Because the images are normalized globally (as opposed to normalizing only in the SC region or only around the surrounding region of the lesion), I am not seeing the benefits of normalization.

Next steps:

• try normalizing only around the lesion region (possibly dilate around the lesion and take the voxels in the dilated region)