Sparse Support Recovery Via Covariance Estimation

Lekshmi Ramesh; Chandra R. Murthy

ICASSP 2018accepted0 citations

Sparse Support Recovery Via Covariance Estimation

Abstract

We consider the problem of recovering the common support of a set of k-sparse signals {x i }L i=1 from noisy linear underdetermined measurements of the form {Φx i + w i }L i=1 where Φ ϵ R m×N (m <; N) is the sensing matrix and wi is the additive noise. We employ a Bayesian setup where we impose a Gaussian prior with zero mean and a common diagonal covariance matrix Γ across all xi, and formulate the support recovery problem as one of covariance estimation. We develop an algorithm to find the approximate maximum-likelihood estimate of Γ using a modified reweighted minimization procedure. Empirically, we find that the proposed algorithm succeeds in exactly recovering the common support with high probability in the k <; m regime with L of the order of m and in the k ≥ m regime with larger L. The key advantage of the proposed algorithm is that its complexity is independent of L, unlike existing sparse support recovery algorithms.

BibTeX

@inproceedings{icassp2018_sparsesupportrec,
  title = {Sparse Support Recovery Via Covariance Estimation},
  author = {Lekshmi Ramesh and Chandra R. Murthy},
  booktitle = {ICASSP 2018},
  year = {2018}
}