[vtkusers] [vtk-developers] Imagedata/Uniformgrid non axis aligned

Will Schroeder will.schroeder at kitware.com
Thu Nov 13 06:20:15 EST 2008


Something similar has been evolving in the ITK world for a while now. While
transforms were added to ITK's image, the problem is that many of the
algorithms are spatially sensitive (e.g., derivatives) and must be rewritten
to accommodate the transform. There is some controversy because this slows
execution down, adds complexity, and the last I checked not everything was
rewritten. But there is a chance that you can do what you want in ITK, and
in case some work has to be done there is a potential path to accept any
changes that you might make. You'll have to deal with interfacing between
VTK and ITK, the more complex coding style of ITK, and of course
contributing changes into ITK can be harder because the community
scrutinizes them more thorougly via the Insight Journal and other processes.

You'll want to check with an ITK expert like Luis Ibanez or Stephen Aylward
(and of course the mailing list) for a more accurate description of the
current ITK situation.


On Thu, Nov 13, 2008 at 3:45 AM, John Biddiscombe <biddisco at cscs.ch> wrote:

> I started writing some code for a particular job yesterday and then
> realized that I can't do what I want to because it is not possible to have a
> non-axis aligned vtkImageData - or better vtkUniformGrid.
> I wonder if there is a solution to having an image-like structure that is
> non axis aligned using some AMR box or heirarchical box dataset instead. I
> had a quick look at the classes, but didn't see an immediate solution. I
> need the dataset to be efficient like vtkImageData, but arbitrarily
> orientated. I need to add a transform to a vtkImageData at the dataset
> level, applying a transform to an actor would not help. The data needs to be
> really transformed in space (for resampling purposes).
> Can anyone suggest an alternative?
> If not, is there any other demand for modifying vtkUniformGrid perhaps to
> take a transform so that functions like findcell/getpoint/bounds etc are
> routed through it? (also, would it be realistic to add this to uniform grid,
> or are there too many places where assumptions are made about voxel corners
> etc).
> thanks
> JB
> --
> John Biddiscombe,                            email:biddisco @ cscs.ch
> http://www.cscs.ch/
> CSCS, Swiss National Supercomputing Centre  | Tel:  +41 (91) 610.82.07
> Via Cantonale, 6928 Manno, Switzerland      | Fax:  +41 (91) 610.82.82
> _______________________________________________
> vtk-developers mailing list
> vtk-developers at vtk.org
> http://www.vtk.org/mailman/listinfo/vtk-developers

William J. Schroeder, PhD
Kitware, Inc.
28 Corporate Drive
Clifton Park, NY 12065
will.schroeder at kitware.com
518-371-3971 (phone and fax)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20081113/956f408d/attachment.htm>

More information about the vtkusers mailing list