This piece of software is provided as support materiel for the article:
M. Moussallam, L. Daudet, G. Richard "Matching Pursuits with Random Sequential Subdictionaries", Signal Processing , 2012
Matching pursuits are a family of greedy algorithms widely used in signal processing to solve sparse approximation and recovery problems. They rely on an atom selection step that requires the calculation of numerous projections, which can be computationally costly for big dictionaries and burdens their competitivity in coding applications.
We propose to use a non adaptive random sequence of subdictionaries in the decomposition process, thus browsing a larger dictionary space in a probabilistic fashion with no additional projection cost nor parameter estimation.