New Upper Bounds of Cumulative Coherence for l1-2-Minimization in Compressed Sensing

This paper mainly studies the compressed sensing
$l$
1-2
-minimization problem and proposes new bounds for cumulative coherence μ
1
(s). We point out that if cumulative coherence μ
1
(s-1) and μ
1
(2s-1) satisfy (2),or cumulative coherence μ
1
(2s-1) satisfies (7) then the sparse signal can can be stably restored in the noise model through the
$l$
l-2
-minimization, and accurately restored in the free noise model.