An efficient augmented Lagrangian method with semismooth Newton solver for total generalized variation

<p style='text-indent:20px;'>Total generalization variation (TGV) is a very powerful and important regularization for various inverse problems computer vision tasks. In this paper, we propose semismooth Newton based augmented Lagrangian method solving problem. The (also called as of multipliers) widely used lots smooth or nonsmooth variational problems. However, its efficiency heavily depends on the corresponding coupled nonlinear system together simultaneously. With efficient primal-dual methods challenging highly subproblems involving total generalized variation, develop competitive compared with some fast first-order method. analysis metric subregularities functions, give both global convergence local linear rate proposed methods.</p>