Exact solutions of one-dimensional total generalized variation (Q2257629)

The scope of this paper is to derive some analytical results regarding the total generalized variation regularization (introduced in [\textit{K. Bredies} et al., SIAM J. Imaging Sci. 3, No. 3, 492--526 (2010; Zbl 1195.49025)]). More precisely, the regularization parameters are investigated in order to find out when is the total generalized variation regularization different from the total variation regularization (i.e. the Rudin-Osher-Fatmi model) and the second order variation model regularization, respectively, and when does it coincide with either of them. Explicit solutions of total generalized variation denoising for particular one-dimensional function data are provided, too. Worth noticing is that one of the main tools for obtaining the mentioned results is represented by the conjugate functions and the Fenchel duality.