[Insight-users] documentation

Zhiyong Xie zyxie98 at yahoo.com
Fri Dec 2 09:49:44 EST 2005


I thought the document of some classes is too simple to give user a clear idea of these classes. It is one reason we miss something which are already implemented. For example, from the document, I have no idea about PolylineMask2DImageFilter for what kind of input it expect and what kind of output we will receive. It will be sad that the contributor spend lot of time to develop a class, but people ignore it only because of insufficient information. So for the contributors, please do not forget to take a little more time to give more detail of stuff you implemented. Your work is greatly appreciated.
   
  Sean   


Zachary Pincus <zpincus at stanford.edu> wrote:
  > 1) The Bresenham line iterator is already in ITK.
> It was contributed by Benjamin King on June 2005
>
> http://www.itk.org/Insight/Doxygen/html/ 
> classitk_1_1LineConstIterator.html
> http://www.itk.org/Insight/Doxygen/html/classitk_1_1LineIterator.html

Hmm, now that I think of it, what's needed here is something more 
like the PathIterator (since Paths can use continuous indices, but 
the LineIterator only takes normal itk::Index points for the line ends).

Fortunately, it seems that a PathIterator is available.

> 2) This iterator is used by the PolylineMask2DImageFilter
> in order to rasterize Polylines into 2D binary mask images.
>
> http://www.itk.org/Insight/Doxygen/html/ 
> classitk_1_1PolylineMask2DImageFilter.html

Great -- I don't know how I missed this.

> 4) What we seem to be missing from your list of desired
> functionalities is a filter for taking a 2D image mask and
> extracting the contours in the form of a PolyLineParametricPath.
>
> However, such filter can easily be written by modifying the
> SimpleContourExtractorImageFilter
>
> http://www.itk.org/Insight/Doxygen/html/ 
> classitk_1_1SimpleContourExtractorImageFilter.html

I'm not fully convinced this will be easy - the tricky part would be 
reconstructing properly-connected poly-line(s) from the individual 
segments generated from a modified contour extractor filter. (Such 
modifications, I think, would basically make it into a "marching 
squares"
filter -- which really isn't hard to implement.) But I suspect there 
will be a not-too-hard way to connect the points properly.

How is point-connection done in the (harder) case of marching cubes? 
Just keeping a list of all the "open" edge points in a mesh to 
connect to?

One other issue is that PolyLines can't capture planar objects with 
'holes' -- you would get extract two PolyLines from a 'donut' shape. 
This really isn't an issue for me since my shapes don't have holes, 
but if I were to make a general contours-to-geometry filter for ITK 
(which I would like to do, if I'm making one for myself!) would this 
issue need to be addressed?

Zach
_______________________________________________
Insight-users mailing list
Insight-users at itk.org
http://www.itk.org/mailman/listinfo/insight-users
  


			
---------------------------------
Yahoo! Shopping
 Find Great Deals on Gifts at Yahoo! Shopping 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20051202/c854bf44/attachment.htm


More information about the Insight-users mailing list