Eric Mootz’s emPolygonizer generates a mesh from the isosurface of a distance field, and while it’s very good at that, in my mind that’s a very different problem from either retopologizing or “joining the dots” of a point cloud.
I had wanted to take a stab at trying to implement power crust or another non-convex remeshing algorithm in ICE as a rainy-day project, but I never got around to it. For those who are interested in the subject, this paper is a few years old but is still a decent overview of the area: http://www.cs.technion.ac.il/~gotsman/AmendedPubl/Pierre/remeshing_survey.pdf Of course you should not expect any such algorithm to have an intuition about which areas will be subject to high deformation, or how to adjust the topology accordingly. gray From: [email protected] [mailto:[email protected]] On Behalf Of Jason S Sent: Tuesday, July 08, 2014 10:15 AM To: [email protected] Subject: Re: Retopo using surface emited particle positions On 07/07/14 23:34, Ed Manning wrote: Well, there are a lot of decent meshers out there, including the built-in one which is based on an older version of Eric Moootz's excellent emPolygonizer. Actually I dont see how (traditionally) polygonized point clouds can have much more useful resulting meshes (topologywise) than an original dense mesh, as opposed to connecting points that have 'guided' positions on a surface where I can definitely see how. The real issue, for reptoplogy to be most useful for animation, is that you need your new topology to have good edgeloop structure and flow as well as variable detail. Much harder! I agree that better results could probably still be achieved by hand, (not unlike UV unwrapping) but I'm also convinced that (more than) decent/acceptable (while quick) results could be automated (not unlike UV pelting) (plus I think it's fair to say that there would be tons of other (shape generating) applications for this, like a more closely fitting point cloud hull) On 07/07/14 22:16, Perry Harovas wrote: Cool idea. I am 100% of no help with this, but I wanted to say I really like where you are going with this. Thanks, I don't know how hard it was to make convex hulls, but it's also beyond my current expertise as well. So perhaps it could serve as an idea for someone that know the ropes. PS; If it may be of help, (Unless Julian or Guillaume would be willing :) both Julian's and Guillaume's ConvexHull source codes are on RRay.de (Guillaumes' is in the same . rar as the compiled) cheers, J
<<attachment: winmail.dat>>
