[Insight-users] 32-bit data from FEM

[Mike Bristow]

Hi,

I am trying to assess the effects of FEM registration by using the 
SquaredDifferenceImageFilter. I seem to be having trouble, however, 
with the 32-bit float output that the FEM registration produces. I am 
trying to read the image data using an ImageFileReader with an input 
pixel type of float, but when I try working with this (rescaling, 
histogram matching, diffing) and outputting this data, the resulting 
files appear corrupted.

Can I get the FEM registration to output something other than float?

Is there something special I should be doing to work with the FEM 
output?

Thanks,
Mike Bristow