1: \begin{abstract}

2: 	\noindent\small 

3: 	We are interested in restoring images having values in a symmetric Hadamard manifold

4: 	by minimizing a functional with a quadratic data term and a total variation like regularizing term.

5: 	To solve the convex minimization problem, we extend the Douglas-Rachford algorithm 

6: 	and its parallel version to symmetric Hadamard manifolds.

7: 	The core of the Douglas-Rachford algorithm are reflections of the functions involved

8: 	in the functional to be minimized.

9: 	In the Euclidean setting the reflections of convex lower semicontinuous functions

10: 	are nonexpansive. 

11: 	As a consequence, convergence results for Krasnoselski-Mann iterations  imply the convergence of the 

12: 	Douglas-Rachford algorithm.

13: 	Unfortunately, this general results does not carry over to Hadamard manifolds, where

14: 	proper convex lower semicontinuous functions can have expansive reflections. 

15: 	However, splitting our restoration functional in an appropriate way, we have 

16: 	only to deal with special functions namely, several distance-like functions

17: 	and an indicator functions of a special convex sets.

18: 	We prove that the reflections of certain distance-like functions on Hadamard manifolds are nonexpansive

19: 	which is an interesting result on its own.

20: 	Furthermore, the reflection of the involved indicator function is nonexpansive

21: 	on Hadamard manifolds with constant curvature so that the Douglas-Rachford algorithm converges here.

22: 		

23: 	Several numerical examples demonstrate the advantageous performance of 

24: 	the suggested algorithm compared to other existing methods

25: 	as the cyclic proximal point algorithm or half-quadratic minimization.

26: 	Numerical convergence is also observed in our experiments 

27: 	on the Hadamard manifold of symmetric positive definite matrices  

28: 	with the affine invariant metric which does not have a constant curvature.	

29: \end{abstract}

30: