用于压缩感知与稀疏去噪的快速线性化Bregman迭代
摘要
我们提出并分析了一种极其快速、高效且简单的方法来求解问题:min{parallel to u parallel to(1) : Au = f, u is an element of R-n}。该方法首次在 [J. Darbon and S. Osher, preprint, 2007] 中描述,更详细的说明见 [W. Yin, S. Osher, D. Goldfarb and J. Darbon, SIAM J. Imaging Sciences, 1(1), 143-168, 2008],严格的理论证明见 [J. Cai, S. Osher and Z. Shen, Math. Comp., to appear, 2008, see also UCLA CAM Report 08-06] 和 [J. Cai, S. Osher and Z. Shen, UCLA CAM Report, 08-52, 2008]。其动机源于压缩感知,该领域如今拥有广泛而令人兴奋的历史,似乎始于 Candes 等人 [E. Candes, J. Romberg and T. Tao, 52(2), 489-509, 2006] 和 Donoho [D. L. Donoho, IEEE Trans. Inform. Theory, 52, 1289-1306, 2006]。大量参考文献见 [W. Yin, S. Osher, D. Goldfarb and J. Darbon, SIAM J. Imaging Sciences 1(1), 143-168, 2008] 以及 [J. Cai, S. Osher and Z. Shen, Math. Comp., to appear, 2008, see also UCLA CAM Report, 08-06] 和 [J. Cai, S. Osher and Z. Shen, UCLA CAM Report, 08-52, 2008]。我们的方法对 [J. Darbon and S. Osher, preprint, 2007] 和 [W. Yin, S. Osher, D. Goldfarb and J. Darbon, SIAM J. Imaging Sciences, 1(1), 143-168, 2008] 中非常高效的方法引入了一项称为“kicking”的改进,并将其应用于欠采样信号的去噪问题。将 Bregman 迭代用于图像去噪始于 [S. Osher, M. Burger, D. Goldfarb, J. Xu and W. Yin, Multiscale Model. Simul, 4(2), 460-489, 2005],并带来了基于全变分方法的改进结果。在此,我们将其应用于信号去噪,特别是本质稀疏的信号,甚至是可能欠采样的信号。
关键词
引用
@article{arxiv.1104.0262,
title = {Fast Linearized Bregman Iteration for Compressive Sensing and Sparse Denoising},
author = {Stanley Osher and Yu Mao and Bin Dong and Wotao Yin},
journal= {arXiv preprint arXiv:1104.0262},
year = {2011}
}