Regular triangulation of irregular mesh.

May 15, 2007 7:20 AM

Given a three dimensional triangular mesh M, composed of irregularly sized triangles, is there a program, plugin¹ or algorithm that will create a new mesh M' with the same overall shape as M but with regularly sized triangles (i.e.:all triangles having similar sizes)? 1: For 3dsMax, Rhino, etc.

I realize that it's not possible to have all the triangles be exactly the same size and could accept a small variation, say about 10-20%. I also realize that some detail of M will be lost but could also accept that.

posted by signal to computers & internet (8 comments total)

Not sure if this is what you are talking about, but in Maya (and possibly other apps) you can "average vertices" repeatedly. Each time it will smooth out the mesh (and deform) but they will become more uniform in size. I don't think this will work though, if the starting mesh has too much variation. If so, I know Lightwave has a reduce polygon tool that will recreate a new mesh based on your parameters. I've only dabbled in 3dsMax so I don't know its particulars.
posted by JimBobNoPants at 8:05 AM on May 15, 2007

JimBobNoPants: most reduce polygon tools try to optimize the surface by curvature, so it creates many different sizes of triangles. Don't know about the Lightwave version, though. The Maya function sounds like it might help, will look into it. Thanks.
posted by signal at 8:11 AM on May 15, 2007

I'm also a heavy Maya user, and JimBobNoPants' answer is the same that I would have given.

JMBNP: Nice to see some other Maya folks out here.
posted by fake at 12:45 PM on May 15, 2007

I tried "average vertcies", and it seems to be a sort of mesh.smooth on steriods, making the entire mesh smaller with each iteration, but not equalizing the triangle sizes. So, I'll keep looking. Can anybody think of an algorithm to do this?
posted by signal at 1:30 PM on May 15, 2007

What do you mean by "same size"? Congruent? Same area? "Small" angles (< 90 degrees, say)?br>
Some of the papers listed here might conceivably be helpful.
posted by leahwrenn at 5:13 PM on May 15, 2007

That is, if you want algorithms...those papers are likely to be pretty theoretical, so won't help if you want something already implemented.
posted by leahwrenn at 5:13 PM on May 15, 2007

leahjwrenn: similar areas. Could you recommend 1 of those papers?
posted by signal at 5:33 PM on May 15, 2007

Good resource, BTW.
posted by signal at 5:35 PM on May 15, 2007