A hybrid overlapping group sparsity denoising model with fractional-order total variation and non-convex regularizer

Although the total variational model can significantly suppress noise and make blurred images sharper, it will unavoidably produce blocky artifacts. In this paper, a new hybrid model based on overlapping group sparsity (OGS) was proposed to effectively prevent the above problems, which combines the advantages of the fractional-order total variational and the non-convex regularizer. The overlapping group sparse fractional-order total variational (OGS-FOTV) regular term can restore complex features in images while reducing noise. Meanwhile, the non-convex regularization term based on the overlapping group sparsity on hyper-Laplacian (OGS-HL) prior can usefully control the staircase artifacts and better protect image edges. To tackle the above denoising model, we employ the alternating direction method of multipliers to decompose our model into several sub-problems and solve them one by one. For images affected by different levels of white Gaussian noise, we conducted a contrast experiment with other advanced methods. The results show that the new model is superior to other related models in denoising.