An Iteratively Reweighted Norm Algorithm for Total Variation Regularization

Total variation (TV) regularization has become a popular method for a wide variety of image restoration problems, including denoising and deconvolution. Recently, a number of authors have noted the advantages, including superior performance with certain non-Gaussian noise, of replacing the standard lscr² data fidelity term with an lscr¹ norm. We propose a simple but very flexible and computationally efficient method, the iteratively reweighted norm algorithm, for minimizing a generalized TV functional which includes both the lscr²-TV and and lscr2-TV problems.