Efficient implementation of matching pursuit using a genetic algorithm in the continuous space

In this work we introduce an alternative implementation of matching pursuit (MP) using a genetic algorithm in the continuous space (GACS). MP is an attractive analysis approach in which the signal is sequentially decomposed into a linear expansion of atoms (functions) from a dictionary of waveforms so as to obtain a sparse representation. The main problem with MP is its computation load, due to the necessarily large size of the dictionary. We propose instead to determine the optimal atom at each stage of the decomposition using a GACS, i.e. a genetic algorithm that requires no quantization of the solution parameters. Preliminary simulation results illustrate the potential benefits of this scheme.

[1]  Sam Kwong,et al.  Genetic Algorithms in Filtering , 1999 .

[2]  Martin Vetterli,et al.  Atomic signal models based on recursive filter banks , 1997, Conference Record of the Thirty-First Asilomar Conference on Signals, Systems and Computers (Cat. No.97CB36136).

[3]  Reto Meuli,et al.  Robust parameter estimation of intensity distributions for brain magnetic resonance images , 1998, IEEE Transactions on Medical Imaging.

[4]  Zhifeng Zhang,et al.  Adaptive Nonlinear Approximations , 1994 .

[5]  B. Rao,et al.  Forward sequential algorithms for best basis selection , 1999 .

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

[7]  Mitsuo Gen,et al.  Genetic algorithms and engineering design , 1997 .

[8]  Zbigniew Michalewicz,et al.  Handling Constraints in Genetic Algorithms , 1991, ICGA.

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

[10]  Shie Qian,et al.  Signal representation using adaptive normalized Gaussian functions , 1994, Signal Process..

[11]  John H. Holland,et al.  Outline for a Logical Theory of Adaptive Systems , 1962, JACM.