A Comparison of Signal Compression Methods by Sparse Solution of Linear Systems

This paper deals with the problem of signal compression by linearly expanding the signal to be compressed along the elements of an overcomplete dictionary. The compression is obtained by selecting a few elements of the dictionary for the expansion. Therefore, signal description is realized by specifying the selected elements of the dictionary as well as their coefficients in the linear expansion. A crucial issue in this approach is the algorithm for selecting, in correspondence of each realization of the signal, the elements of the dictionary to be used for the expansion. In this paper we consider different possible algorithms for basis selection and compare their performances in a practical case of speech signal.

[1]  Stéphane Mallat,et al.  Matching pursuits with time-frequency dictionaries , 1993, IEEE Trans. Signal Process..

[2]  Balas K. Natarajan,et al.  Sparse Approximate Solutions to Linear Systems , 1995, SIAM J. Comput..

[3]  Michael A. Saunders,et al.  Atomic Decomposition by Basis Pursuit , 1998, SIAM J. Sci. Comput..

[4]  Kjersti Engan,et al.  Optimized signal expansions for sparse representation , 2001, IEEE Trans. Signal Process..

[5]  Ronald R. Coifman,et al.  Entropy-based algorithms for best basis selection , 1992, IEEE Trans. Inf. Theory.

[6]  Bhaskar D. Rao,et al.  An affine scaling methodology for best basis selection , 1999, IEEE Trans. Signal Process..

[7]  R. Vanderbei.  LOQO user's manual — version 3.10 , 1999 .