1: \begin{abstract}

2: In this paper, we study the $L_1/L_2$ minimization on the gradient for imaging applications.  Several recent works have demonstrated that  $L_1/L_2$  is better than the $L_1$ norm when approximating the $L_0$ norm to promote sparsity. Consequently, we postulate that applying $L_1/L_2$ on the gradient is better than the classic total variation (the $L_1$ norm on the gradient) to enforce the sparsity of the image gradient. To verify our hypothesis, we consider a constrained formulation to reveal empirical evidence on the superiority of $L_1/L_2$  over $L_1$ when recovering piecewise constant signals from low-frequency measurements. Numerically, we design a specific splitting scheme, under which we can prove   subsequential \tp{and global} convergence for the alternating direction method of multipliers (ADMM) \tp{under certain conditions}. Experimentally, we demonstrate visible improvements of $L_1/L_2$  over $L_1$ and other nonconvex regularizations for image recovery from low-frequency measurements and two medical applications of MRI and CT reconstruction. All the numerical results show the efficiency of our proposed approach.

3: % Motivated by a scale-invariant regularization term	

4: % We consider a type of the gradient-sparse recovery problems, which can be formulated as a $L_0$ minimization. Various kinds of convex/nonconvex terms have been proposed to approximate  $L_0$. Here we focus on the , specifically for the constrained model. Following [C.Wang,et.al 2020], we proposed a similar splitting algorithm for the constrained model. Moreover, under certain conditions, both the subsequential and the global convergence of this approach are provided. We empirically reveal evidences on for recovering pixel-wise constant signals and we apply our method to three different applications. The numerical results shows the efficiency of our proposed approaches.    

5: 	\end{abstract}

6: