[Insight-users] Parameters Tuning -- Help Needed !!!

[Luis Ibanez]

Hi Sateesh,

Maybe I'm missing something in your description,
but what you describe seems to be a normal behavior.

Once you are done with the registration process,
the resulting transform may contain some scaling along
the Z axis. This can easily result in blank slices
being produced at the output of the resampled moving
image.

This can be common since the anatomical region covered
by the CT scan may not be the same as the one covered
by the PET scan.

The blank slices are not a problem in themselves.

What you really need to verify is the right alignment
in the slices that do exist. In other words, you may
have to compare (with a checkerboard, or color encoding,
or with blending...) the fixed image with the resampled
moving image. Ignore the blank slices at the beginning
and end of the volume.

If the two volumes align well the anatomical structures,
you are in good shape an may not need to perform any
further parameter tunning.


Regards,


    Luis


----------------------------------------
cspl wrote:
> Hi Luis,
>     A small problem in Registration. As i said before my data sets are 
> 512x512x205 ( CT images ) and 128x128x205 ( Pet images ). I could 
> register these two data sets according to ur suggestions.
>     Now the problem is i get blank slices in the starting and ending of 
> the registered output. That is 10th slice in registered output is equal 
> to 1st slice in moving data set( Pet) and 200 slice is equal to 205th 
> slice in moving data set. All the slices before 10 and after 200 are 
> blank slices in registered output. Why it is so?
>     Parameters are
>     Translate scale : 10000
>     Max Step Length : 1
>     Min step Lenght : 0.0001
>     Optimizer scale : 1
>     Spatial samples : 1000
>     bins            : 50
>     Levels     : 3
>     Iterations      : 150
>    
> I feel there is a problem with parameters tuning. What values we have to 
> change to get exact output? I changed the translate scale and max step 
> lenght further down but no change in output.
>  
> Please help me in this regard.
>  
> Thanks in advance.
>  
> -Regards,
>    Sateesh.