[vtkusers] Combining green and magenta images to get grey

[DavyW]

Hi there,

I have a question which needs some introduction:
We are writing medical imaging software and we have been using VTK with
great succes in displaying CT/MR scans and on top of that all kinds of
structures. We can register/fuse two image sets by reslicing two scans,
displaying one in normal colours and the other displayed just above the
first one with a transparency and color mapping (using vtkImageMapToColors
and a vtkLookupTable). We can then manually move the transparant image by
just changing the DataOrigin of the vtkImageImport.

Now we want another registration view: mapping the first 3D scan's grey
values to green and the second scan's to magenta. This is no problem either.
However, now we want to 'sum' these (resliced) images so that wherever there
is a good match, green+magenta becomes a grey value. Note that the two 3D
images are arbitrary placed in space and can have different resolution and
size.

I'm aware of things like vtkImageBlend, vtkImageMathematics and so on, but
it's my understanding and experience that these don't account for the image
data's position in space (or even its resolution). 

Our pipeline is like: vtkImageImport -> vtkImageReslice ->
vtkImageMapToColors -> vtkActor to get a '2D' view (with the camera on
parallel projection). And that twice (one for each CT scan). This pipeline
beautifully accounts for the images' size, position and resolution. But
putting for example a vtkImageBlend after the reslicing makes it look like
the vtkImageBlend throws away all knowledge about image position and pretty
much blends the image data regardless of whether the first image set is
placed at (0, 0, 0) and the other at (100, 100, 100).

I hope I have made my problem clear. Any suggestions are welcome!

Kind regards,

Davy


--
View this message in context: http://vtk.1045678.n5.nabble.com/Combining-green-and-magenta-images-to-get-grey-tp5498762p5498762.html
Sent from the VTK - Users mailing list archive at Nabble.com.