• ### Announcements

#### Archived

This topic is now archived and is closed to further replies.

# principle directions

## 6 posts in this topic

Anyone know of a good implementation for calculating principle directions of a manifold with boundary? The only paper I could find was by Taubin, but it'll be nice to find code that works. I would write it myself, but I have one day to do it. I don't think that's possible for me. Perhaps there's a quick and dirty way to calculate this that'll take a really short amount of time? Edited by - davidko on October 10, 2001 12:19:10 AM
0

Timkin
0

##### Share on other sites
How is it specified? you mean, what data structures do I use?

I just have a list of polys and all the polygons it is connected to. Each vertex knows all the polygons that contains it. This information is determined during starttime, but later is discarded. I use the AC3D modeler, which doesn't give you any connectivity information, so I do it myself so I can calculate all the vertex normals and such. I make sure the models I create are manifold.

Edited by - davidko on October 12, 2001 4:08:35 AM
0

##### Share on other sites
And what do you mean by ''principal directions'' ? A generalised manifold can be very complex, so much that it''s impossible to identify it''s direction or directions.
0

##### Share on other sites
http://amp.ece.cmu.edu/Publication/Cha/icip01_Cha.pdf
This one is real easy to understand and implement.

The fast, and obvisously more complex method is here together with source code and example program:
http://www.cs.berkeley.edu/~mirtich/publications.html
"Fast and Accurate Computation of Polyhedral Mass Properties"

Im still trying to find this paper:
Sheue-ling Lien, James T. Kajiya, "A Symbolic Method for Calculating the Intergral Properties of Arbitrary Nonconvex Polyhedra"

Edited by - no way on October 12, 2001 10:02:50 AM
0

##### Share on other sites
The Mirtich reference pointed to by no way is good. I''ve actually used the code and it works, though I don''t recall if it gives you the principal inertias and axes...
0

##### Share on other sites
Ahh...Thanks for the links. I''ll take a look into it.
0