[vtkusers] Marching Cubes Algorithms and Airtight Volume

[Hong Wan]

Hi Will,

 

Thanks for responding to my thread, I have waited days before somebody
actually read my thread and reply me.

 

Yes, you are right that marching cubes will always produce watertight
surfaces, but what if the dataset is incomplete, say dataset of a human
hand that contain 10 slices,

And all  these slices that feed through marching cubes algorithms will
created a water tight surface, what if the last dataset is missing and a
marching cubes is unable to enclose the surfaces?? Is there any other
ways to repair the geometry to enclose it??

 

 

Regards

 

Hong Wan

 

 

-----Original Message-----
From: Will Schroeder [mailto:will.schroeder at kitware.com] 
Sent: Monday, November 11, 2002 8:52 PM
To: Hong Wan
Subject: Re: [vtkusers] Marching Cubes Algorithms and Airtight Volume

 

Marching cubes will ALWAYS produce watertight, manifold surfaces. The
only place there might be "holes" is on the boundary where the surface
exits the dataset. If there are holes then either a boundary exists are
something is going wrong with the segmentation process. If the problem
persists, create an example with a small dataset that demonstrates the
behavior.
Will

At 06:08 PM 11/11/2002 +0800, you wrote:




Hi 

 

Can anyone suggest a step to create an airtight model from marching
cubes algorithms? I need to get rid of the big holes or small holes
presense in the model before exporting to stl format for rapid
prototyping.

 

For example, using Marching Cubes on bones of a human hand, there are
usually areas where segmentation based on thresholding does not do a
good job, instead of creating triangles, it creates holes with leave a
gap in the surface model of the human hand, I am thinking if there are
any ways to fill this gap?? To create a completely enclosed surface
model?

 

Regards

 

Hong Wan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20021112/01a4ea42/attachment.htm>