[Insight-users] danielsson distance map vector field
Luis Ibanez
luis.ibanez@kitware.com
Thu May 13 20:09:48 EDT 2004
Hi David:
1) An example on the use of the DistanceMap for
accelerating the IterativeClosestPoint registration
has been added under
Insight/Examples/Registration/
IterativeClosestPoint3.cxx
2) You don't pass the distance map to the Optimizer
but to the Metric. The IterativeClosestPointMetric
http://www.itk.org/Insight/Doxygen/html/classitk_1_1IterativeClosestPointMetric.html
Already knows how to use the Distance map.
3) You could take the Canny edges and extract a
set of points to be used as a FixedPointSet.
Note that if you follow the example
IterativeClosestPoint3.cxx
you will do something like:
CannyEdges ---> FixedPoints ---> BinaryImage ---> DistanceMap
| |
| |
+------------> ICP Metric <-------+
^
|
|
MovingPoint-----------+
Where there is some redundancy between the Canny edges image
and the Binary image that you feed into the DistanceMap filter.
You could try the following:
CannyEdges -------> DistanceMap ------+
| |
| |
| V
+------- FixedPoints -------> ICPMetric
^
|
|
MovingPoint-----------+
Note that after you do this, you will be registering
the MovingPointSet against the CannyEdges, not against
the FixedPointSet. In that regard, the extraction of
fixed points from the CannyEdges is not that critical....
Regards,
Luis
-------------------------
David Macias Verde wrote:
> Hi all:
>
> How can I introduce the vector field given by the Danielsson Distance
> Map filter into the Levenberg-Marquard optimizer when doing a pointset
> to poinset registration with canny edges and distance map as involved
> images in the process?
>
> Thanks in advance,
>
> David Macias-Verde
>
>
>
> _______________________________________________
> Insight-users mailing list
> Insight-users@itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>
More information about the Insight-users
mailing list