Lecture 8: Mesh Representations and Geometry Processing (74)
atsu-kotani
I was curious if we have an access to a true function of an object surface and the task is now to generate a set of N triangles that best approximate the target surface, how much of violation on good mesh shapes will occur? For a simple shape, such as a sphere, I assume that the optimal solution would be a set of equilateral triangles. Is this goodness of triangular meshes only for the purpose of practical operations to avoid artifacts or does it have anything to do with optimality of mesh representations? Also, if we try to solve this task recursively, from coarse to fine (i.e. N=20, 100, and so on), wouldn't the chance of good meshes increase?
I was curious if we have an access to a true function of an object surface and the task is now to generate a set of N triangles that best approximate the target surface, how much of violation on good mesh shapes will occur? For a simple shape, such as a sphere, I assume that the optimal solution would be a set of equilateral triangles. Is this goodness of triangular meshes only for the purpose of practical operations to avoid artifacts or does it have anything to do with optimality of mesh representations? Also, if we try to solve this task recursively, from coarse to fine (i.e. N=20, 100, and so on), wouldn't the chance of good meshes increase?