Boundedness and unboundedness in total variation regularization

DOI10.1007/S00245-023-10028-YMaRDI QIDQ6170838zbMATH OpenFDO

Authors K. Bredies, José A. Iglesias, Gwenael Mercier

Publication date 10 August 2023

Published in Applied Mathematics and Optimization (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/2203.03264

zbMATH Keywords

boundedness of minimimizersgeneralized taut stringinfimal convolution regularizerslinear inverse problemstotal variationvanishing weights

Mathematics Subject Classification ID

Linear operators and ill-posed problems, regularization (47A52) Variational problems in a geometric measure-theoretic setting (49Q20) Numerical solutions of ill-posed problems in abstract spaces; regularization (65J20) Numerical solution to inverse problems in abstract spaces (65J22)

Abstract: We consider whether minimizers for total variation regularization of linear inverse problems belong to $L^{i}nfty$ even if the measured data does not. We present a simple proof of boundedness of the minimizer for fixed regularization parameter, and derive the existence of uniform bounds for sufficiently small noise under a source condition and adequate a priori parameter choices. To show that such a result cannot be expected for every fidelity term and dimension we compute an explicit radial unbounded minimizer, which is accomplished by proving the equivalence of weighted one-dimensional denoising with a generalized taut string problem. Finally, we discuss the possibility of extending such results to related higher-order regularization functionals, obtaining a positive answer for the infimal convolution of first and second order total variation.

Review text

In mathematics, the total variation identifies several slightly different concepts, related to the local or global structure of the codomain of a function or a measure. The total variation denoising, also known as total variation regularization or total variation filtering, is a noise removal process (filter). It is based on the principle that signals with excessive and possibly spurious detail have high total variation, that is, the integral of the image gradient magnitude is high. The total variation regularization have had an important role in several classical problems of the geometric measure theory, signal processing and physics. Several authors studied the total variation regularization [\textit{V. Caselles} et al., Rev. Mat. Iberoam. 27, No. 1, 233--252 (2011; Zbl 1228.94005); \textit{K. Bredies} et al., SIAM J. Imaging Sci. 3, No. 3, 492--526 (2010; Zbl 1195.49025); \textit{A. Chambolle} and \textit{P-L. Lions}, Numer. Math. 76, No. 2, 167--188 (1997; Zbl 0874.68299); \textit{E. Gonzalez} and \textit{U. Massari} Rend. Sem. Mat. Univ. Politec. Torino 52, No. 1, 1--28 (1994; Zbl 0819.49025); \textit{M. Grasmair} and \textit{A. Obereder} Numer. Funct. Anal. Optim. 29, No. 3--4, 346--361 (2008; Zbl 1142.65013); \textit{M. Grasmair}, J. Math. Imaging Vision 27, No. 1, 59--66 (2007; Zbl 1478.94041); \textit{V. Gutev}, J. Math. Anal. Appl. 491, No. 1, Article ID 124242, 12 p. (2020; Zbl 1519.54006); \textit{C. Kirisits} et al., SIAM J. Imaging Sci. 12, No. 4, 1643--1668 (2019; Zbl 1439.49066); \textit{T. Valkonen}, Inverse Probl. 37, No. 4, Article ID 045010, 30 p. (2021; Zbl 1515.65147)]. The principal objective in this paper is to study boundedness and unboundedness in total variation regularization. The authors present a simple proof of boundedness of the minimizer for fixed regularization parameter, and obtain a boundedness result for the case of infimal convolution of first and second order total variation regularizers, for which the optimality conditions are closely related to subgradients of total variation.

Recommendations

• Regularization of linear least squares problems by total bounded variation
• A note on convergence of solutions of total variation regularized linear inverse problems
• On the regularity of the total variation minimizers
• Structural Properties of Solutions to Total Variation Regularization Problems
• Regularization of linear inverse problems with total generalized variation

Cites work

• A note on convergence of solutions of total variation regularized linear inverse problems
• A notion of total variation depending on a metric with discontinuous coefficients
• An algorithm for mean curvature motion
• An introduction to total variation for image analysis
• Analysis.
• Analytical aspects of spatially adapted total variation regularisation
• Characterization of minimizers of convex regularization functionals
• Comparison of BV norms in weighted Euclidean spaces
• Convergence of level sets in total variation denoising through variational curvatures in unbounded domains
• Convex functions. Constructions, characterizations and counterexamples
• Fine properties of the subdifferential for a class of one-homogeneous functionals
• Fonctions à hessien borné
• Frontiere orientate di curvatura media assegnata in \(L^p\)
• Generalizations of the Taut String Method
• Geometric properties of solutions to the total variation denoising problem
• Higher-order total variation approaches and generalisations
• scientific article; zbMATH DE number 1807400 (Why is no real title available?)
• scientific article; zbMATH DE number 3873956 (Why is no real title available?)
• scientific article; zbMATH DE number 721957 (Why is no real title available?)
• scientific article; zbMATH DE number 1376935 (Why is no real title available?)
• scientific article; zbMATH DE number 1448982 (Why is no real title available?)
• Image recovery via total variation minimization and related problems
• Influence of dimension on the convergence of level-sets in total variation regularization
• Invariant -minimal sets and total variation denoising on graphs
• Lipschitz extensions and approximations
• Minimizing total variation flow
• Nonlinear total variation based noise removal algorithms
• On the density of classes of closed convex sets with pointwise constraints in Sobolev spaces
• On the regularity of the positive part of functions
• Pairings between bounded divergence-measure vector fields and BV functions
• Pairings between measures and bounded functions and compensated compactness
• Parabolic quasilinear equations minimizing linear growth functionals
• Persistence barcodes versus Kolmogorov signatures: detecting modes of one-dimensional signals
• Regularisation, optimisation, subregularity
• Regularity for solutions of the total variation denoising problem
• Regularity results for almost minimal oriented hypersurfaces in R^n
• Regularization methods in Banach spaces.
• Regularization of linear inverse problems with total generalized variation
• Sets of finite perimeter and geometric variational problems. An introduction to geometric measure theory
• Some qualitative properties for the total variation flow
• Some remarks on the staircasing phenomenon in total variation-based image denoising
• The Dirichlet problem for the total variation flow
• The equivalence of the taut string algorithm and BV-regularization
• Total generalized variation
• Tube methods for BV regularization
• Variational Analysis
• Variational methods in imaging
• Weighted BV functions
• Weighted TV minimization and applications to vortex density models

Cited in

(3)

• Aspects of Total Variation RegularizedL¹Function Approximation
• Total variation bound for Milstein scheme without iterated integrals
• Monotonicity of the jump set and jump amplitudes in one-dimensional TV denoising

This page was built for publication: Boundedness and unboundedness in total variation regularization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6170838)