Answers are grounded in this paper and cite their sources.
The first order optimality system of a total variationregularization based variational model with L²-data-fitting inimage denoising (L²-TV problem) can be expressed as an ellipticvariational inequality of the second kind. For a finite elementdiscretization of the variational inequality problem, an aposteriori error residual based error estimator is derived and itsreliability and (partial) efficiency are established. The resultsare applied to solve the L²-TV problem by means of the adaptivefinite element method. The adaptive mesh refinement relies on thenewly derived a posteriori error estimator and on an additional heuristic providing alocal variance estimator to cope with noisy data. The numericalsolution of the discrete problem on each level of refinement isobtained by a superlinearly convergent algorithm based onFenchel-duality and inexact semismooth Newton techniques and whichis stable with respect to noise in the data. Numerical resultsjustifying the advantage of adaptive finite elements solutions arepresented.
No takes yet. Share an insight, caveat, or question.
Hintermüller et al. (2014) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: