 |
Center for Nonlinear Studies |
We introduce a comprehensive framework for mesh improvement for meeting user-desired objectives. Our generic approach translates user objectives into a feasible resampling region, by constraining sample points to lie inside (outside) a collection of simple geometric primitives. We show how to express various mesh quality metrics as geometric constraints and develop a toolbox of sampling-based local operators to improve or preserve quality objectives. We demonstrate the successful application of our framework for various combinations of key metrics such as preserving the minimum angle, reducing the number of Steiner points, and mesh simplification.
Host: Mikhail Shashkov