A nonconvex nonsmooth regularization method with structure tensor total variation

In this paper, a novel regularization method for image restoration and reconstruction is introduced which is accomplished by adopting a nonconvex nonsmooth penalty that depends on the eigenvalues of structure tensor of the underlying image. At first, an alternating minimization scheme is developed in which the problem can be decomposed into three subproblems, two of them are convex and the remaining one is smooth. Then, the convergence of the sequence which generate by the alternating minimization algorithm is proved. Finally, the efficient performance of the proposed method is demonstrated through experimental results for both grayscale and vector-value images.