Automatic Parameterization

There are many mesh processing methods that become much more tractable when we’re provided with a mapping of a complex surface onto a simple domain. The mapping of surfaces into the plane is also of obvious importance in texture mapping.

Required reading

First lecture

Our first topic will focus on conformal mappings of a triangulated surface into the plane.

Plus, you’ll want to read why LSCM=DCP (from David Cohen-Steiner and Mathieu Desbrun).

Second lecture

Next, we’ll be focusing on a slightly different problem. Namely: how can we construct a parameterization of an input mesh onto a simpler base polyhedron? One way of doing this is provided by the algorithm described here:

Third lecture

Finally, here’s an even more general problem. How do we construct smooth mappings of one surface onto another? Mapping onto a simple base polyhedron is now just a special case.

Recommended reading

I highly recommend the fine surveys written by Michael Floater and Kai Hormann. In the context of this course, you’ll probably find the second one to be the most helpful.

The GSP paper is an improvement of the earlier MAPS approach. Reading about the MAPS algorithm might make it easier to understand the GSP algorithm:

To fully appreciate Inter-surface mapping by Schreiner et al., I recommend the following papers:

Another good choice of edge weighting scheme in linear parameterization methods is Floater’s mean value coordinate construction: