Smoothed ℓp-ℓ2 solvers for signal denoising

摘要

The basis pursuit denoising refers to the solution of an ℓ1-ℓ2 minimization formulation which is well known as an effective method for signal denoising. In this paper we investigate an ℓp-ℓ2 formulation with p ∈ (0, 1) for denoising. Based on an analysis of the discontinuity of the global minimizer of the ℓp-ℓ2 problem with respect to regularization parameter, we propose two smoothed ℓp-ℓ2 solvers for orthogonal basis and overcomplete dictionary respectively. Experimental studies that evaluate the performance of the proposed solvers with various parameter settings are also presented.