A fast SVD for multilevel block Hankel matrices with minimal memory storage

Motivated by the Cadzow filtering in seismic data processing, this paper presents a fast SVD method for multilevel block Hankel matrices. A seismic data presented as a multidimensional array is first transformed into a two dimensional multilevel block Hankel (MBH) matrix. Then the Lanczos process is applied to reduce the MBH matrix into a bidiagonal or tridiagonal matrix. Finally, the SVD of the reduced matrix is computed using the twisted factorization method. To achieve high efficiency, we propose a novel fast MBH matrix-vector multiplication method for the Lanczos process. In comparison with existing fast Hankel matrix-vector multiplication methods, our method applies 1-D, instead of multidimensional, FFT and requires minimum storage. Moreover, a partial SVD is performed on the reduced matrix, since complete SVD is not required by the Caszow filtering. Our numerical experiments show that our fast MBH matrix-vector multiplication method significantly improves both the computational cost and storage requirement. Our fast MBH SVD algorithm is particularly efficient for large size multilevel block Hankel matrices.

[1]  I. Dhillon.  Algorithm for the Symmetric Tridiagonal Eigenvalue/Eigenvector Problem , 1998 .

[2]  Wei Xu,et al.  A twisted factorization method for symmetric SVD of a complex symmetric tridiagonal matrix , 2009, Numer. Linear Algebra Appl..

[3]  S. Freire,et al.  Eigenimage Processing of Seismic Sections , 1988 .

[4]  F. Teixeira,et al.  Fast algorithm for matrix–vector multiply of asymmetric multilevel block‐Toeplitz matrices in 3‐D scattering , 2001 .

[5]  K. V. Fernando.  On computing an eigenvector of a tridiagonal matrix , 1997 .

[6]  R. Plemmons,et al.  Structured low rank approximation , 2003 .

[7]  Kenan Y. Sanliturk,et al.  Noise elimination from measured frequency response functions , 2005 .

[8]  Raymond H. Chan,et al.  An Introduction to Iterative Toeplitz Solvers (Fundamentals of Algorithms) , 2007 .

[9]  Golub Gene H. Et.Al.  Matrix Computations, 3rd Edition , 2007 .

[10]  Wei Xu,et al.  Block Lanczos tridiagonalization of complex symmetric matrices , 2005, SPIE Optics + Photonics.

[11]  Jack J. Dongarra,et al.  Lapack95 users' guide , 2001, Software, environments, tools.

[12]  S. Qiao,et al.  A fast symmetric SVD algorithm for square Hankel matrices , 2008 .

[13]  L Tsang,et al.  Scattering of electromagnetic waves from dense distributions of spheroidal particles based on Monte Carlo simulations. , 1998, Journal of the Optical Society of America. A, Optics, image science, and vision.

[14]  Lynn Burroughs,et al.  Prestack Rank-Reduction-Based Noise Suppression , 2009 .

[15]  Gene H. Golub,et al.  Matrix computations , 1983 .

[16]  Liu Yan.  An Improved Method for Noise Reduction Based on Singular Value Decomposition , 2012 .

[17]  M. Sacchi,et al.  A Randomized SVD For Multichannel Singular Spectrum Analysis (MSSA) Noise Attenuation , 2010 .

[18]  Lynn Burroughs,et al.  Rank-Reduction-Based Trace Interpolation , 2010 .

[19]  Michael Ghil,et al.  ADVANCED SPECTRAL METHODS FOR CLIMATIC TIME SERIES , 2002 .

[20]  Daniel B. Szyld,et al.  An introduction to iterative Toeplitz solvers , 2009, Math. Comput..

[21]  G. P. King,et al.  Extracting qualitative dynamics from experimental data , 1986 .

[22]  James A. Cadzow,et al.  Signal enhancement-a composite property mapping algorithm , 1988, IEEE Trans. Acoust. Speech Signal Process..

[23]  Yimin Wei,et al.  A Lanczos bidiagonalization algorithm for Hankel matrices , 2009 .

[24]  Ivan Markovsky,et al.  Low Rank Approximation - Algorithms, Implementation, Applications , 2018, Communications and Control Engineering.

[25]  C. Loan.  Computational Frameworks for the Fast Fourier Transform , 1992 .

[26]  Per-Gunnar Martinsson,et al.  Randomized algorithms for the low-rank approximation of matrices , 2007, Proceedings of the National Academy of Sciences.

[27]  Gene H. Golub,et al.  Matrix computations (3rd ed.) , 1996 .

[28]  M. Sacchi,et al.  Simultaneous seismic data denoising and reconstruction via multichannel singular spectrum analysis , 2011 .

[29]  Stewart Trickett,et al.  F-xy Cadzow Noise Suppression , 2008 .

[31]  Clifford H. Thurber,et al.  Estimating the model resolution matrix for large seismic tomography problems based on Lanczos bidiagonalization with partial reorthogonalization , 2007 .

[32]  Mauricio D. Sacchi,et al.  A fast reduced-rank interpolation method for prestack seismic volumes that depend on four spatial dimensions , 2013 .