1: \begin{abstract}

2: 	%Set also locally

3: 	\parindent0mm

4: 	\parskip0mm

5: 	\noindent

6: In many image and signal processing applications, as interferometric synthetic

7: aperture radar (SAR) or color image restoration in HSV or LCh spaces the data has

8: its range on the one-dimensional sphere $\mathbb S^1$.

9: Although the minimization of total variation (TV) regularized functionals is

10: among the most popular methods for edge-preserving image restoration such

11: methods were only very recently applied to cyclic structures.

12: However, as for Euclidean data, TV regularized variational methods

13: suffer from the so called staircasing effect. This effect can be avoided

14: by involving higher order derivatives into the functional.

15: 

16: This is the first paper which uses higher order differences of cyclic data

17: in regularization terms of energy functionals for image restoration.

18: We introduce absolute higher order differences for

19: $\mathbb S^1$-valued data in a sound way which

20: is independent of the chosen representation system on the circle.

21: Our absolute cyclic first order difference is just the

22: geodesic distance between points. Similar to the geodesic distances the

23: absolute cyclic second order differences have only values in $[0,\pi]$.

24: We update the cyclic variational TV approach by our new cyclic second order

25: differences. To minimize the corresponding functional we apply a cyclic

26: proximal point method which was recently successfully proposed for Hadamard

27: manifolds. Choosing appropriate cycles this algorithm can be implemented in an

28: efficient way. The main steps require the evaluation of proximal mappings of

29: our cyclic differences for which we provide analytical expressions.

30: Under certain conditions we prove the convergence of our algorithm.

31: Various numerical examples with artificial as well as real-world data demonstrate the

32: advantageous performance of our algorithm.

33: \end{abstract}

34: