[Insight-users] 3D affine registration.

[Mostafa Jabarouti Moghaddam]

Hi Luis,

As you know, sometimes the orientation information in the header file is not
correct.
For a clean registration, I do this:
1) I set the origin of both images to (0,0,0)
2) by using a translation transform, I search the space to align the images.
3) I set the center of rotation to the center of the fixed image.
4) using the some rotations, i calculate a similarity measure for choosing
the best rotation.
5) Initialize the transform and then .....


Is it effective? or is there other ways to do that?

Best,
MJM

On Mon, Jan 4, 2010 at 12:40 AM, Luis Ibanez <luis.ibanez at kitware.com>wrote:

> Hi MJM,
>
> The optimization process is very unlikely to correct
> for rotations larger than 20 degrees.
>
> If you know that your images are misaligned by
> more than 20 degrees, then your best option is to
> provide a good initialization for the transform.
>
> For example:
> via the use of the LandmarkTransformInitializer class.
>
>
>
> Beware, however, that it is common for medical images
> to be acquired at different orientations and for those
> orientations to be annotated as part of the image header,
> particularly if you are using DICOM.  Therefore, before
> you conclude that your images are off by as much as 180
> degrees, you should review if there is any orientation
> information in the header of those images.
>
> One good test is to open both images into Slicer.
> Slicer is one of the few applications that will display
> correctly the relative positions of medical images,
> including both origin and orientation.
>
> You can get Slicer from:
> http://www.slicer.org/
>
>
>      Regards,
>
>
>             Luis
>
>
>
> -----------------------------------------------------------------
> On Fri, Dec 25, 2009 at 8:30 AM, Mostafa Jabarouti Moghaddam
> <zavata.afnan at gmail.com> wrote:
> > Hello all,
> >
> > If the orientations of the fixed image and moving image are not same, How
> > can we expand the parameter space? for example, we could change the
> rotation
> > parameter range between [-180,180] in each direction!
> > I searched most of the e-mails in ITK, but i just found out that the best
> > way is choosing the proper initial transform. Unfortunately, choosing the
> > proper initial transform needs an extensive computation!
> >
> >
> > Best,
> > MJM
> >
> > _____________________________________
> > Powered by www.kitware.com
> >
> > Visit other Kitware open-source projects at
> > http://www.kitware.com/opensource/opensource.html
> >
> > Kitware offers ITK Training Courses, for more information visit:
> > http://www.kitware.com/products/protraining.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/20100104/71f9fa6f/attachment-0001.htm>