Open dpar39 opened 1 year ago
Hi, I think this is a misunderstanding of how intrinsic triangulations work. The intrinsic geometry of the original surface is indeed exactly preserved, in the sense that length/angles/areas/etc are unchanged. In this sense you are working with an abstract mesh along the surface of the input geometry. However, if you take the vertex positions and draw straight lines between them, you get a different surface back---as you've noticed here, the corners get cut off.
You could use the common subdivision to extract a mesh which preserves the geometry exactly, but that is generally not what you want.
Usually the whole point with intrinsic triangulations is to avoid realizing the actual intrinsic mesh sitting in space. You generally can't (without approximations like you see here), and if you do you are just back to doing ordinary mesh operations.
I believe it's about replicating section 4.9 Comparison to Traditional Remeshing in Geometry Processing with Intrinsic Triangulations
I was also wondering if there was an easy way to get the remeshed surface but it looks like it must be done by hand.
I'm trying to compute a new mesh from the intrinsic triangulation, but results are not quite as expected. Any clue why some triangles near right edges are being modified like this? Below is the code I'm using to generate a new mesh on which the geodesic path are all surface points on vertices.