1: \begin{abstract}

2:         We are interested in the restoration of noisy and blurry images where the texture

3:         mainly follows a single direction (i.e., directional images). Problems of this type arise, for example,

4:         in microscopy or computed tomography for carbon or glass fibres. {In order to deal with these problems,

5:         the Directional Total Generalized Variation (DTGV) was developed by Kongskov et al.~in 2017 and 2019,

6:         in the case of impulse and Gaussian noise.}

7:         In this article we focus on images corrupted by Poisson noise,

8:         extending the DTGV regularization to image restoration models where the data fitting term is the generalized

9:         Kullback--Leibler divergence. We also propose a technique for the identification of the main texture direction,

10:         which improves upon the techniques used in the aforementioned work about DTGV. We solve the problem by an

11:         ADMM algorithm with proven convergence and subproblems that can be solved exactly at a low computational cost.

12:         Numerical results on both phantom and real images demonstrate the effectiveness of our approach.\\[10pt]

13:         {\textbf{Keywords:} directional image restoration; Poisson noise; DTGV regularization; ADMM method}

14:     \end{abstract}

15: