Exact Recovery Conditions for Sparse Representations with Partial Support Information

TL;DR

This paper establishes new coherence-based conditions for the exact recovery of sparse signals when partial support information is available, improving understanding of recovery guarantees for lp-relaxation, OMP, and OLS methods.

Contribution

It introduces a novel coherence condition for sparse recovery with partial support knowledge, extending classical results and analyzing relationships between NSP and ERC in this context.

Findings

Derived a new coherence condition mu<1/(2k-g+b-1) for informed support recovery.

Showed the coherence condition is independent of restricted-isometry conditions.

Analyzed the relationships between NSP and ERC in the informed sparse recovery setting.

Abstract

We address the exact recovery of a k-sparse vector in the noiseless setting when some partial information on the support is available. This partial information takes the form of either a subset of the true support or an approximate subset including wrong atoms as well. We derive a new sufficient and worst-case necessary (in some sense) condition for the success of some procedures based on lp-relaxation, Orthogonal Matching Pursuit (OMP) and Orthogonal Least Squares (OLS). Our result is based on the coherence "mu" of the dictionary and relaxes the well-known condition mu<1/(2k-1) ensuring the recovery of any k-sparse vector in the non-informed setup. It reads mu<1/(2k-g+b-1) when the informed support is composed of g good atoms and b wrong atoms. We emphasize that our condition is complementary to some restricted-isometry based conditions by showing that none of them implies the other. Because this mutual coherence condition is common to all procedures, we carry out a finer analysis based on the Null Space Property (NSP) and the Exact Recovery Condition (ERC). Connections are established regarding the characterization of lp-relaxation procedures and OMP in the informed setup. First, we emphasize that the truncated NSP enjoys an ordering property when p is decreased. Second, the partial ERC for OMP (ERC-OMP) implies in turn the truncated NSP for the informed l1 problem, and the truncated NSP for p<1.