[MINC-users] mincmath

[Andrew Janke]

Hi Caroline,

My guess is that discretisation of your mask volume is causing the
issue. Meaning instead of 0 being stored a very small number is
instead. What datatype is your mask volume and how are you determining
that the values in the mask are zero?

If you want mask calculations to be robust to this issue then you can
use minccalc to treat voxels with a mask value of less than 0.5 as 0.


   $ minccalc -expression "if(A[1] < 0.5) ? 0 : A[0]" input.mnc
mask.mnc output.mnc


a


On 23 May 2016 at 09:32, Caroline Paquette <caroline.paquette at mcgill.ca> wrote:
> Hello,
> We are using mincmath to multiply a volume (PET acquired with HRRT) by a
> brain mask, which should be simple. Some outputs are non-zero outside the
> brain (although these same voxels are zero in the mask file!). Error is
> random as some of the files multiplied by the same mask are zero while
> others are non-zero (e.g., 3 trials for the same subject, images are
> realigned so we use the same brain mask).
> Any advice? Is this a known problem? We tried writing the files float and
> using different combinations of file formats without success. Files were
> converted from ecat (original) to nifti in matlab using the spm_ecattonii
> function. Preprocessing was done with SPM12 and then converted to minc with
> nii2mnc (we also tried some options with the import without success).
> Thank you,
> Caroline
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users