Bilevel training schemes in imaging for total-variation-type functionals with convex integrands

Abstract

In the context of image processing, given a $k$-th order, homogeneous and linear differential operator with constant coefficients, we study a class of variational problems whose regularizing terms depend on the operator. Precisely, the regularizers are integrals of spatially inhomogeneous integrands with convex dependence on the differential operator applied to the image function. The setting is made rigorous by means of the theory of Radon measures and of suitable function spaces modeled on $BV$. We prove the lower semicontinuity of the functionals at stake and existence of minimizers for the corresponding variational problems. Then, we embed the latter into a bilevel scheme in order to automatically compute the space-dependent regularization parameters, thus allowing for good flexibility and preservation of details in the reconstructed image. We establish existence of optima for the scheme and we finally substantiate its feasibility by numerical examples in image denoising. The cases that we treat are Huber versions of the first and second order total variation with both the Huber and the regularization parameter being spatially dependent. Notably the spatially dependent version of second order total variation produces high quality reconstructions when compared to regularizations of similar type, and the introduction of the spatially dependent Huber parameter leads to a further enhancement of the image details.

arXiv comment27 pages, 6 figures

References in corpus (10)

• A combined first and second order variational approach for image reconstruction
• Multi-Contrast MRI Reconstruction with Structure-Guided Total Variation
• Bilevel parameter learning for higher-order total variation regularisation models
• On Rank-One Convex Functions that are homogeneous of Degree One
• The structure of optimal parameters for image restoration problems
• Higher-order total variation approaches and generalisations
• Embeddings for $\mathbb{A}$-weakly differentiable functions on domains
• On the Trace Operator for Functions of Bounded $\mathbb{A}$-Variation
• A function space framework for structural total variation regularization with applications in inverse problems
• Dualization and Automatic Distributed Parameter Selection of Total Generalized Variation via Bilevel Optimization

Cited by in corpus (2)

• Dyadic partition-based training schemes for TV/TGV denoising
• Bounded variation spaces with generalized Orlicz growth related to image denoising