Sparse Coding Algorithm with Negentropy and Weighted ℓ1-Norm for Signal Reconstruction

摘要: Compressive sensing theory has attracted widespread attention in recent years and sparse signal reconstruction been widely used processing communication. This paper addresses the problem of recovery especially with non-Gaussian noise. The main contribution this is proposal an algorithm where negentropy reweighted schemes represent core approach to solution problem. formalized as a constrained minimization problem, objective function sum measurement error statistical characteristic term, negentropy, regularization lp-norm, for 0 < p 1. however, leads non-convex optimization which difficult solve efficiently. Herein we treat lp -norm serious weighted l1-norms so that sub-problems become convex. We propose optimized combines forward-backward splitting. fast succeeds exactly recovering signals Gaussian Several numerical experiments comparisons demonstrate superiority proposed algorithm.