A Convergent Fixed-Point Proximity Algorithm Accelerated by FISTA for the ℓ0 Sparse Recovery Problem

We propose an approximation model of the original l0 minimization model arising from various sparse signal recovery problems. The objective function of the proposed model uses the Moreau envelope of the l0 norm to promote the sparsity of the signal in a tight framelet system . This leads to a non-convex optimization problem involved the l0 norm. We identify a local minimizer of the proposed non-convex optimization problem with a global minimizer of a related convex optimization problem. Based on this identification, we develop a two stage algorithm for solving the proposed non-convex optimization problem and study its convergence. Moreover, we show that FISTA can be employed to speed up the convergence rate of the proposed algorithm to reach the optimal convergence rate of \(\mathcal {O}(1/k^2)\). We present numerical results to confirm the theoretical estimate.

[1]  Jianqing Fan,et al.  Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties , 2001 .

[2]  Pablo A. Parrilo,et al.  Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization , 2007, SIAM Rev..

[3]  Stephen P. Boyd,et al.  Enhancing Sparsity by Reweighted ℓ1 Minimization , 2007, 0711.1612.

[4]  Wotao Yin,et al.  Improved Iteratively Reweighted Least Squares for Unconstrained Smoothed 퓁q Minimization , 2013, SIAM J. Numer. Anal..

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

[6]  H. Zou The Adaptive Lasso and Its Oracle Properties , 2006 .

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

[8]  C. Micchelli,et al.  Proximity algorithms for image models: denoising , 2011 .

[9]  Lixin Shen,et al.  Proximity algorithms for the L1/TV image denoising model , 2011, Advances in Computational Mathematics.

[10]  Emmanuel J. Candès,et al.  The Power of Convex Relaxation: Near-Optimal Matrix Completion , 2009, IEEE Transactions on Information Theory.

[11]  O. Mangasarian Minimum-support solutions of polyhedral concave programs * , 1999 .

[12]  D. Donoho For most large underdetermined systems of linear equations the minimal 𝓁1‐norm solution is also the sparsest solution , 2006 .

[13]  Karen Uhlenbeck,et al.  Bäcklund transformations and loop group actions , 1998, math/9805074.

[14]  Stephen P. Boyd,et al.  Log-det heuristic for matrix rank minimization with applications to Hankel and Euclidean distance matrices , 2003, Proceedings of the 2003 American Control Conference, 2003..

[15]  Tony F. Chan,et al.  Total Variation Wavelet Inpainting , 2006, Journal of Mathematical Imaging and Vision.

[16]  Yuesheng Xu,et al.  Wavelet inpainting with the ℓ0 sparse regularization , 2016 .

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

[18]  Mila Nikolova,et al.  Analysis of the Recovery of Edges in Images and Signals by Minimizing Nonconvex Regularized Least-Squares , 2005, Multiscale Model. Simul..

[19]  Lixin Shen,et al.  The moreau envelope approach for the l1/TV image denoising model , 2014 .

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

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

[22]  Lixin Shen,et al.  Multi-step fixed-point proximity algorithms for solving a class of optimization problems arising from image processing , 2014, Advances in Computational Mathematics.

[23]  I. Daubechies,et al.  Iteratively reweighted least squares minimization for sparse recovery , 2008, 0807.0575.

[24]  Christian Jutten,et al.  Recovery of Low-Rank Matrices Under Affine Constraints via a Smoothed Rank Function , 2013, IEEE Transactions on Signal Processing.

[25]  S. Foucart,et al.  Sparsest solutions of underdetermined linear systems via ℓq-minimization for 0 , 2009 .