[Insight-users] GroupSpatialObject

[Kajetan Berlinger]

Hello,

I want to realize within a 2D-2D registration application, that only some 
regions (blobs) are used for metric computation. Therefore, per blob I 
created an object of type itkBlobSpatialObject and added it to an object of  
type itkGroupSpatialObject, in order to have one spatial object, which 
contains all the information. I set this spatial object by using  
metric->SetFixedImageMask(groupSpatialObject) to the current metric object.
But in every iteration 0 pixels are considered for computing the similarity 
measure.

I think the problem is, that in the metrics computation classes (e.g. 
NormalizedCorrelationImageToImageMetric) the IsIndide(..) method of the 
spatial object is called without taking into account its depth. And therefore 
the added blobs are never considered as points of the spatial object, because 
the depth is by default 0.

Don't know if this is a bug or if there is something wrong the way I want to 
realize it?

Many thanks for your help!
Kaj