ReRecovery of sparse signals via generalized orthogonal matching pursuit: A new analysis

Abstract

As an extension of orthogonal matching pursuit (OMP) for improving the recovery performance of sparse signals, generalized OMP (gOMP) has recently been studied in the literature. In this paper, we present a new analysis of the gOMP algorithm using the restricted isometry property (RIP). We show that if a measurement matrix Phi is an element of R-mxn satisfies the RIP with isometry constant delta(max{9,S+1}K) <= 1/8, then gOMP performs stable reconstruction of all K-sparse signals x is an element of R-n from the noisy measurements y = Phi x + v, within max {K, left perpendicular8K/Sright perpendicular}iterations, where is the noise vector and S is the number of indices chosen in each iteration of the gOMP algorithm. For Gaussian random measurements, our result indicates that the number of required measurements is essentially m = O (K log n/K), which is a significant improvement over the existing result m = O (K-2 log n/K), especially for large K.