[Insight-developers] TCon topics (repost)

[Damion Shelton]

Hello...

The forward of this message from George came through as an attachment, so
here it is in unattached form.

-------------

I would like to propose the following topics for discussion tomorrow:

1) "Should images (and meshes) know how big they are?"

With itkPhysicalImage, the image knows its origin, spacing, number of
dimensions, transform, etc. but doesn't appear to know how big it is. This
info is accessible via
itkPhysicalImage::GetLargestPossibleRegion()::GetSize() but this seems a bit
counterintuitive. The (as yet unresolved) problem I was having yesterday,
with the GetLargestPossibleRegion()::GetSize() call within
ImageToImageFilter::GenerateData() returning a size of {0,0,0, got me
thinking about how image size is handled.

Perhaps (hopefully?) I'm just missing something obvious.

2) "How do we know who's done / is doing / will do what?"

I'm going back and re-examining what I'm doing in light of Vikram's
suggestion about itkImageFunction. I think this raises the larger question
of how we tell what capabilities ITK has. I see a few options:

a) An ongoing forum for posting present capabilities, current work, and
planned work
b) Emailing the list every time something new is started, to avoid
redundancy
c) Not worrying about it ;-)

In any case, despite the dramatic improvement in naming conventions, it's
still not always clear where to look in the class hierarchy to see whether
what you're interested in exists or not. For instance, I missed
itkImageFunction because I was looking under
itkImageSource/itkImageToImageFilter under the assumption that such a thing
would be part of the process pipeline.

Anyways, thoughts are welcome.

-Damion-