[MINC-users] making slices using mincreshape

[Mishkin Derakhshan]

Hi Lisa,
I've ran into image range problems and data type conversion problems
in the past so now I do as Andrew suggested, but I found that it also
helps to make sure your minc image is a float too:

mincresample -float orig.mnc orig_float.mnc
mnc2nii -float orig_float.mnc orig.nii

mishkin




On Fri, Feb 10, 2012 at 2:00 PM, Lisa F. Akiyama <lrisa87 at uw.edu> wrote:
> Hi Andrew and Gang,
>
> Thank you for your advice.
> Yes, what both of you had suspected was the case and I was able to create
> slice files with the original intensity range intact.
>
> Is there an option in mnc2nii to preserve the original intensity range when
> converting from MINC to NIFTI (or ANALYZE)?
> When I try to convert a MINC file that has a range of 0-100 (according to
> mincinfo) to NIFTI, I am getting a NIFTI while with a mere range of 0-2.
>
>
> Thank you.
>
>
> Best,
> Lisa
>
>
>
> On Thu, Feb 9, 2012 at 4:18 PM, Gang Liang <gang.liang.2011 at gmail.com>wrote:
>
>> Andrew is right.
>>
>> mincresample and mincreshape do not always keep the original range. to keep
>> the original range, use this option:
>>
>> mincresample -keep_real_range
>> or
>> mincreshape -valid_range min max
>> where you can get min and max using mincinfo or mincheader minc_file|grep
>> range
>>
>> Hope this help.
>>
>>
>> On Thu, Feb 9, 2012 at 5:51 PM, Andrew Janke <a.janke at gmail.com> wrote:
>>
>> > Hi Lisa,
>> >
>> > On 10 February 2012 06:39, Lisa F. Akiyama <lrisa87 at uw.edu> wrote:
>> > > I have been using mincresample to slice up MINC images into individual
>> > > axial MINC slice files using the following command options:
>> > > mincreshape -clobber -dimrange zspace=${slice},1 orig.mnc
>> > orig_${slice}.mnc;
>> > >
>> > > However, I am getting slices with a lot of white noise on slices that
>> > > should be completely black according to what I see
>> > > in that slice location on the original image. I get these odd outputs
>> for
>> > > axial slices superior of the head/skull.
>> > > These issues aren't present for black slices containing no tissue
>> > inferior
>> > > of the head.
>> >
>> > My guess would be that there is indeed noise in these slices (with no
>> > real data). as part of its operation mincreshape will set the min and
>> > max of an output image (slice) to the range of the data in it.
>> >
>> > So, if your image data ranges between 0 and 5000 and the noise in the
>> > background between 0 and 10 then a slice in the middle will have an
>> > output range of 0-5000 and you won't see the noise much as it is well
>> > down in the colourmap of tools like register/Display.
>> >
>> > If instead you pick a slice position with no data then the output
>> > range will be 0-10 and you will see the noise now. So, I'd be checking
>> > the output range of your slices with white noise in them. (mincstats
>> > will do).
>> >
>> > This also might not be the case!, if not a screenshot of what is going
>> > on will help.
>> >
>> > Thanks
>> >
>> >
>> > a
>> > _______________________________________________
>> > MINC-users at bic.mni.mcgill.ca
>> > http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>> >
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users