harefort

04-17-2009, 12:47 PM

Hey everybody!

I want to create a "relative topology mapper" script which can transfer vertex/edge/face-data between meshes with identical edge-flow, but different vertex-positions.

(Think of two different heads that were created from the same basemesh)

The thing is that I don't know where to start.

One way would be to ask the user to select the two faces that are identical in the two meshes. Then from there on one would have to create a relative topology map that stores all necessary connections of vertices/edges/faces. If these maps are identical in both objects, one can copy data from one to the other, like UVs, material IDs, smoothing groups...

Another way does everything on it's own, without any user input beyond specifying the two objects. This one would have to find a starting face, create the map for each object and and then compare these for identity.

The important thing is that it is completely position idependant (ideally), so that even in local object space vertices can move freely.

That's where I'm stuck.

Does anybody have experience with something like this, or can direct me to some articles/papers about this topic?

Thanks a ton!

I want to create a "relative topology mapper" script which can transfer vertex/edge/face-data between meshes with identical edge-flow, but different vertex-positions.

(Think of two different heads that were created from the same basemesh)

The thing is that I don't know where to start.

One way would be to ask the user to select the two faces that are identical in the two meshes. Then from there on one would have to create a relative topology map that stores all necessary connections of vertices/edges/faces. If these maps are identical in both objects, one can copy data from one to the other, like UVs, material IDs, smoothing groups...

Another way does everything on it's own, without any user input beyond specifying the two objects. This one would have to find a starting face, create the map for each object and and then compare these for identity.

The important thing is that it is completely position idependant (ideally), so that even in local object space vertices can move freely.

That's where I'm stuck.

Does anybody have experience with something like this, or can direct me to some articles/papers about this topic?

Thanks a ton!