A Comparison of Numerical Schemes for Dictionary Iterative Refinement in Problems of Sparse Approximation

Lukovenkova O.O., Marapulets Yu.V., Tristanov A.B., Kim A.A.
  1. Institute of Cosmophysical Research and Radio Wave Propagation FEB RAS

Sparse approximation is a decomposition of the original signal into the sum of the minimum possible number of functions selected from the given dictionary. Matching pursuit is an algorithm that allows to find an estimated solution to the sparse approximation problem. It has cubic time complexity with respect to the size of the dictionary, its functions length and the signal length. The paper is devoted to the description and comparison of numerical methods that allow to optimize the matching pursuit algorithm in terms of time and memory. The analysis of geoacoustic emission signals using the matching pursuit modifications was carried out.

O.O.Lukovenkova, Yu.V.Marapulets, A.B.Tristanov, A.A.Kim. A Comparison of Numerical Schemes for Dictionary Iterative Refinement in Problems of Sparse Approximation // Proceedings of 2018 XXI IEEE International Conference on Soft Computing and Measurments Scopus SCM. 2018.