Sparse Approximation by Linear Programming: Measuring the Error with the $ell_1$ Norm

In this report we study the problem of sparse signal approximation over redundant dictionaries. We focus our attention on the minimization of a cost function where the error is measured using a l1 norm. We show a constructive equivalence between this minimization and Linear Programming. A recovery condition is then proved and finally we provide an example of the use of such a technique for denoising.