Original Post
I'm trying to implement mesh parameterization using harmonic mapping in order to generate uv coords for a given mesh. I'm trying to follow the following power point presentation. http://www2.in.tu-clausthal.de/~hormann/parameterization/BarycentricMappings2008.ppt I'm using a fixed boundary method where I'm mapping all boundary vertices to the unit circle. I've got this done already, so all uv coordinates for these boundary vertices are already found. Now I have to find the uv coords for internal vertices. I'm having a hard time understanding how to set up the linear equations involved. A * U = U' A * V = V' Where A contains all the computed weight coefficients for each vertex, U and V are the actual 2d coordinates we're trying to find. And then U', V' are i have no idea. I know that barycentric coordinates requires that: SUMMATION ( wij * ui ) = 0 for all neighbors j of i So that would make me think that U' should be all zero? How do the boundary vertex values come into play here? Also the correct way, once I've filled in U' and V', to solve for U and V, is simply: U = At * U' V = At * V' Where At is A transpose. Is this correct? It should be directly solvable, right?