1: \begin{abstract}

2: We consider the thresholding scheme, a time discretization for mean curvature flow

3: introduced by Merriman, Bence and Osher. 

4: We prove a convergence result in the multi-phase case.

5: The result establishes convergence towards a weak formulation of mean curvature

6: flow in the BV-framework of sets of finite perimeter.

7: The proof is based on the interpretation of the thresholding scheme

8: as a minimizing movements scheme by Esedo\u{g}lu et.\ al..

9: This interpretation means that the thresholding scheme preserves

10: the structure of (multi-phase) mean curvature flow as a gradient flow 

11: w.\ r.\ t.\ the total interfacial energy. 

12: More precisely, the thresholding scheme is a minimizing movements scheme for an

13: energy functional that $\Gamma$-converges to the total interfacial energy.

14: In this sense, our proof is similar to the convergence results of Almgren, Taylor and Wang and

15: Luckhaus and Sturzenhecker, which establish convergence of a more academic

16: minimizing movements scheme. Like the one of Luckhaus and Sturzenhecker, 

17: ours is a {\it conditional} convergence

18: result, which means that we have to assume that the

19: time-integrated energy of the approximation converges to

20: the time-integrated energy of the limit. This is a natural assumption, which however

21: is not ensured by the compactness coming from the basic estimates.

22: 

23: \medskip

24: 

25: \noindent \textbf{Keywords:} Mean curvature flow, Thresholding, MBO scheme, Minimizing movements

26: 

27: \medskip

28: 

29: 

30: \noindent \textbf{Mathematical Subject Classification:} 35A15, 65M12

31: \end{abstract}

32: