Multiresolution Parameter Choice Method for Total Variation Regularized Tomography

Abstract

A computational method is introduced for choosing the regularization parameter for total variation (TV) regularization. The approach is based on computing reconstructions at a few different resolutions and various values of regularization parameter. The chosen parameter is the smallest one resulting in approximately discretization-invariant TV norms of the reconstructions. The method is tested with X-ray tomography data measured from a walnut and compared to the S-curve method. The proposed method seems to automatically adapt to the desired resolution and noise level, and it yields useful results in the tests. The results are comparable to those of the S-curve method; however, the S-curve method needs a priori information about the sparsity of the unknown, while the proposed method does not need any a priori information (apart from the choice of a desired resolution). Mathematical analysis is presented for (partial) understanding of the properties of the proposed parameter choice method. It is rigorously proven that the TV norms of the reconstructions converge with any choice of regularization parameter.

A Semiblind Regularization Algorithm for Inverse Problems with Application to Image Deblurring

“…The discussion on how to choose the appropriate α E , α TV , and β is out of the scope of this paper and thus we choose the optimal one, i.e., the one which gives the highest value of SNR among some tested ones. For a discussion on how to select the regularization parameter for the TV regularization see, e.g., [34,28,31,14,17,19,26,13,32,33]. Moreover, for a discussion on multiparameter selection see, e.g., [20].…”

Identifying conductivity in electrical impedance tomography with total variation regularization

“…Although there exists a large number of papers on the numerical solution of the inverse problems of EIT, among these also papers considering the Kohn-Vogelius functional (see, e.g., [28,29]) and total variation regularization (see, e.g., [21,36]), we have not yet found investigations on the discretization error in a combination of both functionals for the fully nonlinear setting, a fact which motivated the research presented in this paper.…”

Controlled wavelet domain sparsity for x-ray tomography

“…Many approaches for the regularization parameter selection have been proposed. For a selection of methods designed for total variation (TV) regularization see the following studies: [5,6,7,8,9,10,11,12]. In this paper we introduce a novel automatic method for choosing µ based on a control algorithm driving the sparsity of the reconstruction to an a priori known ratio 0 ≤ C pr ≤ 1 of nonzero wavelet coefficients in f .…”