[Insight-users] Reading 3D Binary Volumes

[Luis Ibanez]

Hi Aya,

In order to advice you,

we first need you to tell us what is an:

             ItkSFDataImage



      Thanks,

              Luis


--------------------------------------------------------------------
On Thu, Oct 22, 2009 at 12:23 PM, Aya El Gebeely
<aya.elgebeely at symbyo.com> wrote:
> Hi Luis,
>
> Thanks for your reply .. it was very  beneficial
>
> Actually we have our own pipeline for reading a DICOM directory and loading
> into memory.. but let me ask you another question if don't mind :
>
> Using our pipeline we already have in memory the whole volume pixel data as
> well as ,  DICOM info ..
>
> Actually I've a pointer to the volume data in memory ..
>
> How could I load  the volume data into an ItkSFDataImage and its data as
> well , to be able to apply various  ITK filters on it?
>
>
> Regards,
> Aya R. ElGebeely
> Software Engineer
> Symbyo Technologies
>
>
> 2009/10/22 Luis Ibanez <luis.ibanez at kitware.com>
>>
>> Hi Aya,
>>
>> The simplest way of loading a "raw" binary file into ITK is
>> to write a MetaImage header for it.
>>
>> This is simply a text header file, that you will put in the same
>> directory as your binary file.
>>
>> As described in:
>> http://www.itk.org/ITK/resources/links.html
>>
>> The minimal MetaImage header would be a file with extension
>> ".mhd", named for example:
>>
>>                             "image.mhd"
>>
>> with content such as:
>>
>> NDims = 3
>> DimSize = 181 217 181
>> ElementType = MET_UCHAR
>> ElementSpacing = 1.2 1.2 3.4
>> Offset = 13.4 27.5 43.2
>> ElementByteOrderMSB = False
>> ElementDataFile = myBinaryFile.raw
>>
>>
>> Where, of course, you will have to replace the
>> "myBinaryFile.raw" with the actual filename of
>> your binary file.
>>
>> The DimSize field with the actual number of pixels
>> along each dimension of your image.
>>
>> The ElementType with the pixel type of the data.
>> Presumably you are doing CT or MRI, and in that
>> case you want to use the pixel type MET_SHORT.
>>
>> Offset should have the coordinates of the image
>> origin.
>>
>> All that said,
>> Please note that you could could have just read
>> the DICOM image directly into ITK, as explained
>> in the ITK Software Guide
>>
>>      http://www.itk.org/ItkSoftwareGuide.pdf
>>
>> in the "Reading and Writing Images" Chapter.
>>
>> You will find source code examples on how to
>> read DICOM  images in the directory:
>>
>>            Insight/Examples/IO
>>
>>
>> The other options that you mention are also valid.
>>
>> You could read the image as DICOM and save it
>> as a .vtk legacy file (although in this case the
>> direction cosines will be lost... but you already
>> lost them anyways in your current process of
>> saving only the binary data).
>>
>> You could also save the image as an Analyze
>> or Nifti image file format. But please note that
>> Analyze only supports three image orientations.
>>
>> My suggestion will be to save the image using
>> the MetaImage file format, or to keep it as a
>> DICOM file.  Those are the two options that
>> will keep the largest amount of information about
>> your image.
>>
>>
>>    Regards,
>>
>>
>>           Luis
>>
>>
>> ------------------------------------------------------------------------------
>> On Thu, Oct 22, 2009 at 10:26 AM, Aya El Gebeely
>> <aya.elgebeely at symbyo.com> wrote:
>> > Dear Luis,
>> >
>> ...
>> >
>> > I 've a question concerning  reading 3D images & applying ITK filters on
>> > 3D
>> > images.
>> >
>> > I have 3D volumes that were constructed from DICOM series & were saved
>> > in
>> > Binary Files. [the pixel data]
>> >
>> > Here is the issue :
>> >
>> > I need to reconstruct the volume, and apply some ITK filters on it such
>> > as
>> > segmentation, Diluting, Laplacian and others
>> >
>> > I'm facing a problem in reading the binary file using ItkImageReader, it
>> > fails to load pixel data.
>> >
>> > - Should I use another kind of reader with the binary volume ?
>> > - Should I stop using binary file and save the volume in another format
>> > ?
>> >
>> > What is your opinion about this problem ?
>> >
>> > Thanks in advance and waiting for your reply :)
>> >
>> > Best Regards,
>> > Aya R. ElGebeely
>> > Software Engineer
>> > Symbyo Technologies
>> >
>
>