On recovery of block sparse signals via block generalized orthogonal matching pursuit

As a greedy algorithm for compressed sensing reconstruction, generalized orthogonal matching pursuit (gOMP) has been proposed by Wang et al. [10] to recover sparse signals, which simply selects multiple indices without additional postprocessing operation. In this paper, we consider efficient method for the recovery of sparse signals that exhibit additional structure in the form of the non-zero coefficients occurring in clusters. A block version of gOMP is proposed, named block generalized orthogonal matching pursuit (BgOMP). Moreover, theoretical analysis based on restricted isometry property (RIP) for BgOMP is investigated. Simulation results show that BgOMP has considerable recovery performance comparable to state-of-the-art algorithms in terms of probability of exact reconstruction and running time.