[Insight-users] Registration (itk Plugin for volview)

[Luis Ibanez]

Hi Robert,

Thanks for letting us know that the new VolView
plugin works for you now.

Note that for the test you are doing by eroding one
of the images it is to expect that the registration
may give you shifted results, since now the image may
fit in many different positions and still be inside
the projection of the other one. There are therefore
multiple locations that will give you similar metric
values.

The optimizer will certainly fall in the closest
local minima. If the initial set of paramters is
inside the capture region of the local minima, then
a gradient descent will lead you there. You could
avoid local minima by using Evolutionary optimizers
at the price of evaluating the metric a larger number
of times. You could also help your registration algorihtm
by providing 3DMarkers (Landmarks) to be used as an initial
guess for the registration. It depends on how many image do
you expect to process in a day, and how much expertise the
operator has.

One interesting exercise you can make is to sweep the
parametric space by selecting different angles.
...  but,... that will just be useful if you wantto
confirm the hypotesis above, and may raise questions
in your advisor's mind regarding the purpose of your
current activities   :-)

If your application really requires to fit one image
inside an eroded one, then you need to come up with a
customized image metric that takes such characteristic
into account. You could also be more pragmatic and actually
erode  your fixed image before registration in order to make
it more similar to the moving image. The erosion filter may
be included as part of the preprocessing in the Image
Registration plugin. This will be easier than creating an
entirely new image metric.



Please let us know what option work best for you,



   Thanks


      Luis


------------------------
Atwood, Robert C wrote:

> Dear Luis:
> 
> Thank you, and whoever else worked on fixing the bug; the registration
> works perfectly on the same test cases I tried before, including the
> combined erosion and skew-axis rotation case. With 3 erosions and a
> rotation of [ 0.58 -0.58 0.58 ] 0.4 it looks a little funny but it is
> probably converging on a valid local minimum ; there is quite a lot of
> overlap. 
>  
> 
> If the initial orientation is close, is it more likely to converge on
> the correct registration; i.e. is the optimizer more likely to find the
> local minimum near to the starting location in the search space?
> 
> Robert
> 
> 
> 
> 
> 
> 
> 
>   
> 
> -----Original Message-----
> From: Luis Ibanez [mailto:luis.ibanez at kitware.com] 
> Sent: 07 July 2004 02:36
> To: Atwood, Robert C
> Cc: Insight-users at itk.org
> Subject: Re: [Insight-users] Registration (itk Plugin for volview)
> 
> 
> 
> 
> Hi Robert,
> 
> 
> A bug on the Jacobian computation of the itkVersorRigid3DTransform was
> fixed last week.  New versions of the VolView plugins have been
> generated after the bug was fixed.
> 
> 
> You can download the DLLs of this new plugins from
> 
>              public.kitware.com
> 
>            /pub/itk/VolViewPlugins
> 
> 
> Please let us know if this new version of the plugins work for
> registering your images.
> 
> 
>    Thanks
> 
> 
> 
>        Luis
> 
> 
> ------------------------
> Atwood, Robert C wrote:
> 
> 
>>Dear Luis:
>>First, I am not sure how to determine the version number of a plugin 
>>DLL downloaded from public.kitware.com. However, I have two versions 
>>which I will cal the 'Old' and the 'New'.
>>
>>The Old appears to have the date '11/12/2003', and the New was 
>>downloaded from public.kitware.com yesterday and appears to have the 
>>date '040518' on that site (though I am not sure how far to trust 
>>windows-reported dates ;-)
>>
>>The shrinking behaviour occurs with the Old but so far not with the 
>>New, however the initial registration is much better with the Old than
> 
> 
>>it is with the New.
>>
>>... <snip>
> 
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>