Fast Nonoverlapping Block Jacobi Method for the Dual Rudin-Osher-Fatemi Model

We consider nonoverlapping domain decomposition methods for the Rudin--Osher--Fatemi~(ROF) model, which is one of the standard models in mathematical image processing. The image domain is partitioned into rectangular subdomains and local problems in subdomains are solved in parallel. Local problems can adopt existing state-of-the-art solvers for the ROF model. We show that the nonoverlapping relaxed block Jacobi method for a dual formulation of the ROF model has the $O(1/n)$ convergence rate of the energy functional, where $n$ is the number of iterations. Moreover, by exploiting the forward-backward splitting structure of the method, we propose an accelerated version whose convergence rate is $O(1/n^2)$. The proposed method converges faster than existing domain decomposition methods both theoretically and practically, while the main computational cost of each iteration remains the same. We also provide the dependence of the convergence rates of the block Jacobi methods on the image size and the number of subdomains. Numerical results for comparisons with existing methods are presented.

[1]  L. Bregman.  The relaxation method of finding the common point of convex sets and its application to the solution of problems in convex programming , 1967 .

[2]  Marc Teboulle,et al.  A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems , 2009, SIAM J. Imaging Sci..

[3]  Jongho Park,et al.  Domain Decomposition Methods Using Dual Conversion for the Total Variation Minimization with L1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{ , 2018, Journal of Scientific Computing.

[4]  Michael Hintermüller,et al.  Non-Overlapping Domain Decomposition Methods For Dual Total Variation Based Image Denoising , 2014, Journal of Scientific Computing.

[5]  Carola-Bibiane Schönlieb,et al.  Bregmanized Domain Decomposition for Image Restoration , 2013, J. Sci. Comput..

[6]  Patrick L. Combettes,et al.  Signal Recovery by Proximal Forward-Backward Splitting , 2005, Multiscale Model. Simul..

[7]  A. Chambolle,et al.  A remark on accelerated block coordinate descent for computing the proximity operators of a sum of convex functions , 2015 .

[8]  Marc Teboulle,et al.  Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems , 2009, IEEE Transactions on Image Processing.

[9]  Amir Beck,et al.  On the Convergence of Block Coordinate Descent Type Methods , 2013, SIAM J. Optim..

[10]  Michael Hintermüller,et al.  Subspace Correction Methods for a Class of Nonsmooth and Nonadditive Convex Variational Problems with Mixed L1/L2 Data-Fidelity in Image Processing , 2013, SIAM J. Imaging Sci..

[11]  Carola-Bibiane Schönlieb,et al.  A convergent overlapping domain decomposition method for total variation minimization , 2009, Numerische Mathematik.

[12]  Xue-Cheng Tai,et al.  Augmented Lagrangian Method, Dual Methods, and Split Bregman Iteration for ROF, Vectorial TV, and High Order Models , 2010, SIAM J. Imaging Sci..

[13]  Antonin Chambolle,et al.  A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging , 2011, Journal of Mathematical Imaging and Vision.

[14]  Jongho Park,et al.  A Finite Element Approach for the Dual Rudin-Osher-Fatemi Model and Its Nonoverlapping Domain Decomposition Methods , 2018, SIAM J. Sci. Comput..

[15]  Tom Goldstein,et al.  The Split Bregman Method for L1-Regularized Problems , 2009, SIAM J. Imaging Sci..

[16]  Carola-Bibiane Schönlieb,et al.  Subspace Correction Methods for Total Variation and 1-Minimization , 2007, SIAM J. Numer. Anal..

[17]  R. Tyrrell Rockafellar,et al.  Convex Analysis , 1970, Princeton Landmarks in Mathematics and Physics.

[18]  Andrea Toselli,et al.  Domain decomposition methods : algorithms and theory , 2005 .

[19]  ANTONIN CHAMBOLLE,et al.  An Algorithm for Total Variation Minimization and Applications , 2004, Journal of Mathematical Imaging and Vision.

[20]  Junfeng Yang,et al.  A New Alternating Minimization Algorithm for Total Variation Image Reconstruction , 2008, SIAM J. Imaging Sci..

[21]  Hyenkyun Woo,et al.  Block Decomposition Methods for Total Variation by Primal–Dual Stitching , 2016, J. Sci. Comput..

[22]  Marc Teboulle,et al.  On the rate of convergence of the proximal alternating linearized minimization algorithm for convex problems , 2016, EURO J. Comput. Optim..

[23]  Chang-Ock Lee,et al.  Primal Domain Decomposition Methods for the Total Variation Minimization, Based on Dual Decomposition , 2017, SIAM J. Sci. Comput..

[24]  Danping Yang,et al.  Convergence Rate of Overlapping Domain Decomposition Methods for the Rudin-Osher-Fatemi Model Based on a Dual Formulation , 2015, SIAM J. Imaging Sci..

[25]  Barry Smith,et al.  Domain Decomposition Methods for Partial Differential Equations , 1997 .

[26]  L. Rudin,et al.  Nonlinear total variation based noise removal algorithms , 1992 .

[27]  P. Tseng.  Convergence of a Block Coordinate Descent Method for Nondifferentiable Minimization , 2001 .