[Insight-users] Any plans for including Chamfer Matching for Registration ?

[Luis Ibanez]

Hi George,

We don't have any plans for including Chamfer Matching
as a metric for registration.  The closest thing available
in ITK is the NarrowBandToImageRegistration method
described in the following paper

http://www.itk.org/cgi-bin/viewcvs.cgi/Papers/MedIASpecialIssue/NarrowBandToImageRegistration/NarrowBandToImageRegistration.pdf?rev=1.1&root=InsightDocuments&view=log

This method generates a cloud of points in a narrow band
around the edges of an anatomical object, and then proceeds
to register the cloud of points to another image.


If you are motivated to implement the Chamfer Matching
as a registration technique, we will be more than happy
to assist you in the process and to help you get that
method into the toolkit.



Please let us know.


    Thanks


      Luis



----------------------------
Li, George (NIH/NCI) wrote:

> Luis:
> 
> Is there any plan to include an early registration
> method called chamfer matching into ITK? Mutual
> information sometimes posts reliability issue on
> the registration, especially for those images that
> have some anatomical changes, such as surgical 
> removal of tumor and thereafter patient weight loss.
> 
> The chamfer matching, a surface-based registration,
> rather than voxel-based, should be superior in this 
> particular case, mostly for CT and MR images from
> brain cancer therapy.
> 
> It was a popular registration method back to ~1993.
> 
> Regards,
> 
> George
> 
>