Print

---

Hi,

here is a (not so polished) list of converters and DICOM header
parsers, along with my personal experience with those tools.

The structure of the list is as follows:
url
  purpose of the program
    name of the program
      my personal experience

http://surfer.nmr.mgh.harvard.edu/fswiki/mri_convert
    converter
        mri_convert: good. must pick random file from series, will
automatically find the remaining files

http://www.cabiatl.com/mricro/mricron/dcm2nii.html
    converter
        dcm2nii
            buggy
                segfault when src path ends with '/'
                does sometimes not read the orientation, slice
thickness or pixel spacing correctly, and does not allow for
correction on the command line or elsewhere
                    does NOT issue a warning that some values could
not be read from the header, which can result in left/right flips
            reads default values from config file, which are
overwritten by command line
            one can not choose the output filename, one can only
choose between some predefined sprintf strings
    overview on bottom

http://godzilla.kennedykrieger.org/dicomfiles/
    converter
        dcm_convert: Philips ACS/NT (not yet tested)
        solaris binary only
    overview

http://cbi.nyu.edu/software/dinifti.php
    converter
        dinifti
            based on ctn, ctn is very good
            has man pages
            is not able to read number of slices, allows for correction
            can not read orientation
            source code looks unstable (c strtok, round from float to
integer to determine number of slices)

http://erl.wustl.edu/research/dicom/ctn.html
    header
        dcm_dump_file: (very good)
        dicomhead: (could not read anything)

dcmtk
    header
        dcmdump: (works)

libgdcm-tools
    header
        gdcmdump
    converter
        gdcmconv: (dcm -> dcm)

http://lcni.uoregon.edu/~jolinda/MRIConvert/
    converter
        mcverter
            segfaulted
        MRIconvert (gui)

http://xmedcon.sourceforge.net
    converter nifti -> dicom
        medcon: cli
        xmedcon: gui
        works for nifti -> dicom
        dicom -> nifti untested

http://www.loni.ucla.edu/Software/Debabeler
    converter
        debabeler
            can not read time series, concatenates all volumes in the
z direction (see views->dicom 2 nifti translation->image dimensions)
            sets all dimensions starting with pixdim5 to zero, which
causes fslview to exit with a failing assertion (would expect 1
instead of 0)

oro.dicom, oro.nifti

ezDICOM


Cheers,
Klaus

On Tue, Sep 25, 2012 at 8:05 PM, Chuck Theobald <[log in to unmask]> wrote:
> Hello Basile,
>
> You could try MRIConvert: http://lcni.uoregon.edu/~jolinda/MRIConvert/
>
> We respond to support queries and code often, even in the midst of a move
> across campus to a new Center. Note that the package comes with a
> command-line program, mcverter, to support scripted processing pipelines.
>
> Regards,
> Chuck
>
>
>
> On 9/25/12 8:45 AM, Basile Pinsard wrote:
>>
>> Dear all
>>
>> I post here questions about dcm2nii because this sofware despite it's
>> widespread use does not have any kind of support (no list/no working email
>> address...).
>> There seems to be regular updates and it's working but the documentation
>> is not updated and options that have been listed on the webpage are not
>> functional such as BeginClip and LastClip or UINT16toFLOAT32 and a lots of
>> the options added by default in dcm2nii.ini are not documented.
>>
>> How many of you are using it in "production", how do you deal with this
>> limited support?
>> Are there any other good alternatives for fmri dicom to nifti conversion
>> that work on all constructors data?
>>
>> Thanks for your thought about what I consider to be a problem in the
>> development of reliable and maintainable processing pipelines.
>>
>> Regards.
>>
>
>
> --
> Chuck Theobald
> System Administrator
> The Robert and Beverly Lewis Center for Neuroimaging
> University of Oregon
> P: 541-346-0343
> F: 541-346-0345