An improved method for OMP-based algorithm; using fusing strategy

Abstract

The Orthogonal Matching Pursuit (OMP) algorithm has been intensively applied for its simple structure and reconstruction efficiency. However, the reconstruction accuracy of OMP algorithm still needs to be improved. Though the OMP-based algorithms, LAOMP algorithm and KOMP algorithm proposed in recent years have significantly improved reconstruction performance while it can lead to fairly high computational complexity. In this paper, we propose an improved method for OMP-based algorithms using fusing strategy and demonstrate its superiority. We use the double intersection method to fuse the estimated support set from algorithms with low computational complexity to get an atomic set with high accuracy. This is used as an initial support set to improve the LAOMP algorithm and KOMP algorithm. According to the simulation results, compared with the original algorithms, the reconstruction accuracy of the improved algorithms increases to some extent while the reconstruction time sharply decreases and the performance is more remarkable as the computational amount of the original algorithms increases.