[Insight-users] Help needed on Metric values !!!!
Luis Ibanez
luis.ibanez@kitware.com
Sat, 08 Feb 2003 07:16:07 -0500
Hi Sateesh,
Could you please be more specific about the
initial parameters you added to the registration
when you passed from 2D to 3D ?
While the registration is running, do you get
all the print out messages on the console ?
the ones that show the iteration number ?
What seems to be vulnerable in your registration
is the low number of Z slices. A volume of size
256 x 256 x 3 is certainly very flat.
What are the spacings in this image ?
In Z is probably much larger that X,Y....
Please try the following:
1) Don't use multiresolution with this data
Subsampling in this volume is uncertain.
The ImagePyramid works by first blurring
your image with a gaussian and then decimating
the resolution. Doing that in a disproportionate
volume may not produce a reasonable pyramid.
e.g. What factors are you using for decimating
at each level of the pyramid ? ...
2) Try one of the Non-multiresolution examples
and plot the outputs provided by the printout
messages. That will allow you to track the
path of the optimizer in the parametric space.
Please let us know what you find.
Thanks
Luis
-------------------------------------------
boggarapu sateesh kumar wrote:
> Hi Lius,
>
> I could register 2d images using the second
> example of MultiResImageRegistration2.cxx. I could
> able to tune the parameters with ur suggestion given
> before. Thanks for the all the inputs. It is working
> fine for rotation, translation, scaling.
>
> Now i have shifted to 3d image registration. I am
> using the first example of imageregistration1.cxx to
> translate an moving image (translated in (30,30,0)). I
> changed the dimensions from 2 to 3. I have added one
> more initial parameters to registration. That's it no
> other changes in code. I executed the application but
> i could not get the output. It is just terminating. It
> is not giving any itk or any other error but jut
> terminating after executing the startregistration
> function. What could be the reason??. My data sets are
> 256x256x3 an MRI images. What else i should change in
> the code??
>
> Please help me in this regard.
>
> Thanks in advance.
>
> -Regards,
> Sateesh.
>
> =====
>
>
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
> http://mailplus.yahoo.com
>