It was mentioned that any remaining irregular vertices were from the original mesh. Is there a method of up sampling that does not have to contain the original vertices?
Sybil
wouldn't it be better to contain the original vertices so it has more of the old information? Might be good for when you downsample the mesh and it will look closer to the original
Isaaz
Since quad meshes are better for subdivision, is there any algorithm to convert from triangle meshes to quad meshes?
keenan
@hlin Yes, for the second part of A2 you can implement an isotropic remesher; by setting the target edge length smaller than the original mesh you'll get more nice, regular triangles. You can of course integrate this with subdivision: e.g., apply Loop subdivision to get a nice shape, then perform isotropic remeshing to get nicer triangles.
keenan
@Isaaz Great question. There are "dumb" ways to turn triangles into quads... like apply one round of Catmull-Clark subdivision! But the question of how to get a nice quad mesh from a triangle mesh is a really interesting and challenging one. This survey describes a lot of the methods that have been developed.
It was mentioned that any remaining irregular vertices were from the original mesh. Is there a method of up sampling that does not have to contain the original vertices?
wouldn't it be better to contain the original vertices so it has more of the old information? Might be good for when you downsample the mesh and it will look closer to the original
Since quad meshes are better for subdivision, is there any algorithm to convert from triangle meshes to quad meshes?
@hlin Yes, for the second part of A2 you can implement an isotropic remesher; by setting the target edge length smaller than the original mesh you'll get more nice, regular triangles. You can of course integrate this with subdivision: e.g., apply Loop subdivision to get a nice shape, then perform isotropic remeshing to get nicer triangles.
@Isaaz Great question. There are "dumb" ways to turn triangles into quads... like apply one round of Catmull-Clark subdivision! But the question of how to get a nice quad mesh from a triangle mesh is a really interesting and challenging one. This survey describes a lot of the methods that have been developed.