1: \begin{abstract}

2: 

3: This work aims to explore the regularity properties of 

4: the smoothed-TV regularization for the functions is of

5: the class H\"{o}lder continuous. 

6: Over some compact and convex domain $\Omega,$

7: we study construction of multivariate function

8: $\varphi(\bx) :\Omega \subset \R^{3} \rightarrow \R_{+}$

9: as the optimized solution to the following convex 

10: minimization problem

11: 

12: \bea

13: \argmin_{\Omega} \left\{

14: F_{\alpha}(\cdot , f^{\delta}) := \frac{1}{2} \norm{\cT( \cdot ) - f^{\delta}}_{\cH}^2 + \alpha J(\cdot) \right\} ,

15: \nonumber

16: \eea

17: where the penalizer 

18: $J(\cdot) : \cC^{1}(\Omega,\R^{3})\rightarrow \R_{+}$ is 

19: the smoothed total variation penalizer

20: 

21: \bea

22: J(\cdot) = \int_{\Omega} \sqrt{\norm{\nabla(\cdot)}_2^2 + \beta} d \bx ,

23: \nonumber

24: \eea

25: 

26: \noindent for a fixed $0 < \beta < 1.$ We assume our target function to be H\"{o}lder

27: continuous. With this assumption, we establish relation between total 

28: variation of our target function and its H\"{o}lder coefficient.

29: We prove that the smoothed-TV regularization is an admissible

30: regularization strategy by evaluating the discrepancy 

31: $\norm{\cT\varphi_{\alpha} - f^{\delta}} \leq \tau\delta$

32: for some fixed $\tau \geq 1.$

33: To do so, we need to assume that the target function

34: to be class of $\cC^{1+}(\Omega).$ From here, under the fact that the penalty 

35: $J(\cdot)$ is strongly convex, we move on

36: to showing the convergence of $\norm{\varphi_{\alpha} - \varphi^{\dagger}},$

37: for $\varphi_{\alpha}$ is the optimum and $\varphi^{\dagger}$

38: is the true solution for the given minimization problem above. 

39: We demonstrate that strong convexity

40: and $2-$convexity are actually different names for the same

41: concept. In addition to these facts,

42: we make us of Bregman divergence in order to be able to quantify

43: the rate of convergence.

44: 

45: \bigskip

46: \textbf{Keywords.}

47: {H\"{o}lder continuity, Bounded variation, smoothed total variation, Morozov discrepancy.}

48: \end{abstract}

49: