[Insight-users] GDCM problems

[Frank Miller]

Ranjan, Sohan (GE, Research) wrote:
> Hi Mathieu.
> 
> Thank you for your reply. I downloaded the CVS version of the code, and
> do not see any difference. The MI values computed are all wrong. May I 
> know what changes have been made in your recent update?
> 
> Sohan
> 
> -----Original Message-----
> From: Mathieu Malaterre [mailto:mathieu.malaterre at kitware.com]
> Sent: Thursday, July 13, 2006 10:40 PM
> To: Ranjan, Sohan (GE, Research)
> Cc: insight-users at itk.org
> Subject: Re: [Insight-users] GDCM problems
> 
> 
> Ranjan, Sohan (GE, Research) wrote:
> 
>>Hello.
>>
>>I had written code for registering images in DICOM format. The code worked well with ITK-2.4.
>>However I switched over to ITK-2.6, and then my code went erratic. What I have noticed is
>>that the problem is coming due to origin of the read image. If a DICOM image read by ITK-2.4
>>read the origin as, say (X,Y,Z), the same image origin read by ITK-2.6 reports it as (X, -Y, Z).
>>Note the change in the sign of the y-corodinate. With this change, the corresponding points 
>>in the two image being registered becomes different, thus the whole MI computation goes
>>haywire, and registration fails. 
>>I am sure many would have faced the same problem, so what is the solution?
> 
> 
> Ranjan,
> 
>     If you have a chance, could you try again with ITK CVS. I removed 
> something that has been discussed for a while about this issue. I 
> believe this would solve your problem.
> 
> Thanks
> Mathieu
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users

Ranjan,

How are the images oriented? i.e. what is the value of the (0020,0037) 
Image Orientation (Patient) in the dicom header? If the images are not 
axis-aligned, you should use itk::OrientedImage.

See:

http://public.kitware.com/pipermail/insight-users/2006-April/017436.html

Frank