[Insight-users] question about image and mesh interaction

[wanlin]

Hi, Alexander,

On Sun, Mar 6, 2011 at 2:22 AM, Alexander Taghva <alextaghva at gmail.com>wrote:

> Hello,
> I am relatively new to ITK so I apologize if this is a basic question.  I
> am working on implementing an algorithm to extract the brain from non-brain
> on an MRI.  In order to do this, I am using a deformable mesh model.
> Specifically, I am using RegularSphereMeshSource and expanding it to match
> the image.  There are a few things I am trying to figure out how to do
> efficiently:
>

1. Calculate statistics (e.g. median intensity) for all image points inside
> the mesh.
>
You can use itk::TriangleMeshToBinaryImageFilter to get binary image of your
mesh, then perform statistics operation using
 itk::StatisticsLabelMapFilter. (Not sure the performance, i guest you will
do it in each iteration).

>
> 2. Iterate through the mesh to figure out all points connected to any given
> vertex (when I implemented the algorithm in Matlab/C++ mex files, I built an
> adjacency list for each vertex).
>
> What is your given vertex, is it a voxel? what is the meaning of connection
of points to vertex (if for voxel, is it all points inside a voxel cube)?


> 3. Extract the final brain image (specifically, label all points outside
> the mesh 0) after deformation.
>
>  itk::TriangleMeshToBinaryImageFilter and itk::MaskImageFilter

Sorry for the length of this post.  Thanks for your help.
>
>
>

wanlin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20110306/22ab0436/attachment-0001.htm>