“…Variants and improvements of this approach have been proposed in the past two decades: the squared ℓ 2 -norm in the fidelity term was replaced by the ℓ 1 -norm in [3,4]; in [5,6] the authors proposed a nonlocal TV regularization, while in [7] the authors minimized TV with a nonlocal data fidelity term. In [8] and in [9] the authors proposed a weighted difference of anisotropic and isotropic TV and an Oriented Total variation ℓ1/2 as regularization term, respectively. In [10] the authors presented the spatially adaptive total variation model that applies less smoothing near significant edges by utilizing a spatially varying weight function that is inversely proportional to the magnitude of image derivatives, while in [11] the authors proposed the use of a new edge indicator, named difference curvature, which can effectively distinguish between edges and ramps, in an adaptive TV method.…”