1: \begin{abstract}

2:     We introduce a robust optimization method for flip-free distortion

3:     energies used, for example, in parametrization, deformation, and volume

4:     correspondence.

5:     This method can minimize a variety of distortion energies, such

6:     as the symmetric Dirichlet energy and our new symmetric gradient energy.

7: We identify and exploit the special structure of distortion energies

8:     to employ an operator splitting technique, leading us to propose a novel

9:     \textbf{A}lternating \textbf{D}irection \textbf{M}ethod of

10:     \textbf{M}ultipliers (ADMM) algorithm to deal with the non-convex,

11:     non-smooth nature of distortion energies.

12: The scheme results in an efficient method where the global step involves a single matrix multiplication and the local steps are closed-form per-triangle/per-tetrahedron expressions that are highly parallelizable.

13:     The resulting general-purpose optimization algorithm exhibits 

14:     robustness to flipped triangles and tetrahedra in initial data as well as

15:     during the optimization.

16: We establish the convergence of our proposed algorithm

17:     under certain conditions and demonstrate applications to parametrization, deformation, and volume correspondence.

18: \end{abstract}

19: