[Insight-users] conversion of scan transofmration matrix between different modality

[Bill Lorensen]

Now I understand.

Not all itk supported formats support orientation. Off the top of my head,
the following do:
DICOM (the gdcm implementation GDCMImageIO)
nrrd
meta

For others, the application has to deal with it. There is a class in itk:
OrientImageFilter that will reshuffle a volume into a selected orientation.

Handling medical image data properly is very tricky. We have tried in itk to
the right thing in DICOM, which is the clinical medical imaging standard.
meta and nrrd are great research formats and we think we handle orientation
properly.

If you are working in a clinical environment, I would strongly recommend
that you use DICOM. If you are in a research environment and will not be
providing information to a clinical environment, then I guess you can use
whatever works for your particular scanner protocols.

Bill

On Thu, Oct 1, 2009 at 10:37 AM, kevin cavin <beibeiwen2003 at yahoo.com>wrote:

> Hello, Bill:
>
>         At this moment I retrieve the dicom voume from the MR and CT
> machines and the images are in AVS field format due to internal reasons.
> Then I use ITK to register the MR and CT volume. I display both of voulme
> and find the orientation is different. I manually adjust the orientation and
> get good registration.
>
>         If I understand what you said correctly, if both volume are read in
> dicom format, the registration should take care the difference of
> orientation automatically. If so, It will be necessary to retrieve the voume
> in dicom format.
>
>         But I still do not understand how ITK can deal with issue
> automatically. Does it mean only the volume in DICOM will work in this way.
> If the volumes are saved in other format, such as VTK, AVS etc, the
> application should deal with the orientation transformation.
>
>         More explanation will be appreciated. Thanks ! Since  I am using
> python instead of C++ to programm, I do not know whether  print command will
> work in the same way.
>
> Calvin
>
>
>
>
>
>
>
>
>
>
> --- On *Thu, 10/1/09, Bill Lorensen <bill.lorensen at gmail.com>* wrote:
>
>
> From: Bill Lorensen <bill.lorensen at gmail.com>
> Subject: Re: [Insight-users] conversion of scan transofmration matrix
> between different modality
> To: "kevin cavin" <beibeiwen2003 at yahoo.com>
> Cc: insight-users at itk.org
> Date: Thursday, October 1, 2009, 9:11 PM
>
>
> Calvin,
>
> As I said, the orientation difference should be taken care of during the
> registration process. Is the data being read as DICOM? If not, how are you
> reading it?
>
> Please ->Print(std::cout) the fixed and moving images after you read them
> and post the results of the Print.
>
> Bill
>
> On Thu, Oct 1, 2009 at 8:31 AM, kevin cavin <beibeiwen2003 at yahoo.com<http://us.mc1120.mail.yahoo.com/mc/compose?to=beibeiwen2003@yahoo.com>
> > wrote:
>
>>   Hello, Bill :
>>
>>      Thanks for  your reply.  What I hve found with the orginal volume is
>> the orientation for CT and MRI scan is different. If I directly register
>> these  two volume, I will get wrong registration result.  I assume the
>> transformtion of orientation for one scan is necessary. but how can I do
>> that with available ITK ?
>>      I am using ITK version 3.12.
>>
>> Kind regards
>>
>> Calvin
>>
>>
>> --- On *Wed, 9/30/09, Bill Lorensen <bill.lorensen at gmail.com<http://us.mc1120.mail.yahoo.com/mc/compose?to=bill.lorensen@gmail.com>
>> >* wrote:
>>
>>
>> From: Bill Lorensen <bill.lorensen at gmail.com<http://us.mc1120.mail.yahoo.com/mc/compose?to=bill.lorensen@gmail.com>
>> >
>> Subject: Re: [Insight-users] conversion of scan transofmration matrix
>> between different modality
>> To: "kevin cavin" <beibeiwen2003 at yahoo.com<http://us.mc1120.mail.yahoo.com/mc/compose?to=beibeiwen2003@yahoo.com>
>> >
>> Cc: insight-users at itk.org<http://us.mc1120.mail.yahoo.com/mc/compose?to=insight-users@itk.org>
>> Date: Wednesday, September 30, 2009, 11:36 PM
>>
>>
>> ITK should handle this. ITK uses the information stored in DICOM to
>> compute an origin, spacing and direction. You should not have to convert the
>> volume, assuming the patient scans are landmarked (in the scanner) at a
>> similar location.
>>
>> What version of ITK are you using?
>>
>> On Wed, Sep 30, 2009 at 11:04 AM, kevin cavin <beibeiwen2003 at yahoo.com<http://us.mc1120.mail.yahoo.com/mc/compose?to=beibeiwen2003@yahoo.com>
>> > wrote:
>>
>>>
>>> Hi, Folks:
>>>
>>>       When I use ITK to register two dicom volume  from different
>>> modalities.  I have found that MR and CT scan often have different scan
>>> transformation matrixes. For example for CT scan the z direction is the y
>>> direction of MR scan. This will pose problem to register the two volumes
>>> with ITK. I am wondering whether in ITK the probelm can be solved soneway by
>>> using some function embedded. Do I need to write individual program to
>>> manula convert one of the volume.  Thanks in advance !
>>>
>>> Kind regards
>>>
>>> Calvin
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> _____________________________________
>>> Powered by www.kitware.com
>>>
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>>
>>> Please keep messages on-topic and check the ITK FAQ at:
>>> http://www.itk.org/Wiki/ITK_FAQ
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.itk.org/mailman/listinfo/insight-users
>>>
>>>
>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20091001/a1d56b64/attachment.htm>