Research reports

Years: 2025 2024 2023 2022 2021 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991

Total Variation Regularization by Iteratively Reweighted Least Squares on Hadamard Spaces and the Sphere

by P. Grohs and M. Sprecher

(Report number 2014-39)

Abstract
We consider the problem of reconstructing an image from noisy and/or incomplete data, where the image/data take values in a metric space \(X\) (e.g. \(\mathbb{R}\) for grayscale, \(S^2\) for the chromaticity component of RGB-images or \(SPD(3)\), the set of positive definite \(3\times 3\)-Matrices, for Diffusion Tensor Magnetic Resonance Imaging (DT-MRI)). We use the common technique of minimizing a total variation (TV) functional \(J\). After having defined \(J\) for arbitrary metric spaces \(X\) we will propose an adaption of the Iteratively Reweighted Least Squares (IRLS) algorithm to minimize \(J\). For the case of \(X\) being a Hadamard space, such as \(SPD(n)\), we prove existence and uniqueness of a minimizer of a regularized functional \(J^\epsilon\) where \(\epsilon>0\) and show that these minimizers convergence to a minimizer of \(J\) when the regularization parameter \(\epsilon\) tends to zero. We show that IRLS can also be applied for \(X\) being a half-sphere. For the case of \(X\) being a Riemannian manifold we propose to use Newton's method on Manifolds to numerically compute the minimizer of \(J^\epsilon\). To demonstrate our algorithm we present some numerical experiments where we denoise and/or inpaint sphere-valued and SPD-valued images.

Keywords: Iteratively reweighted least squares, total variation, regularization, manifold-valued data

@Techreport{GS14_589,
  author = {P. Grohs and M. Sprecher},
  title = {Total Variation Regularization by Iteratively Reweighted Least Squares on Hadamard Spaces and the Sphere},
  institution = {Seminar for Applied Mathematics, ETH Z{\"u}rich},
  number = {2014-39},
  address = {Switzerland},
  url = {https://www.sam.math.ethz.ch/sam_reports/reports_final/reports2014/2014-39.pdf },
  year = {2014}
}

Disclaimer
© Copyright for documents on this server remains with the authors. Copies of these documents made by electronic or mechanical means including information storage and retrieval systems, may only be employed for personal use. The administrators respectfully request that authors inform them when any paper is published to avoid copyright infringement. Note that unauthorised copying of copyright material is illegal and may lead to prosecution. Neither the administrators nor the Seminar for Applied Mathematics (SAM) accept any liability in this respect. The most recent version of a SAM report may differ in formatting and style from published journal version. Do reference the published version if possible (see SAM Publications).