Beyond coherence: Recovering structured time-frequency representations (Q2470546): Difference between revisions

The authors consider the problem of recovering a structured sparse representation of a signal in a overcomplete time-frequency dictionary with a particular structure. For infinite dictionaries that are the union of a wavelet basis and a Wilson basis, sufficient conditions are given for the basis pursuit (\(l^1\)-minimization) and (orthogonal) matching pursuit algorithms (also known as greedy algorithms) to recover a structured representation of an admissible signal. The conditions allow very large (even infinite) support sets to be recovered.