|
This paper demonstrates theoretically that if the restrictedisometry constant of the compressed sensing matrixsatisfies a sufficient condition,then a greedy algorithm called Orthogonal Matching Pursuit (OMP) canrecover a signal with K nonzero entries in K iterations. Incontrast, matrices are also constructed with restricted isometryconstant satisfying a stronger conditionsuch that OMP can not recover K-sparse x in K iterations. Thisresult shows that the conjecture given by Dai and Milenkovic is true. |
|
Keywords:Function theory, compressed sensing, restricted isometry property, orthogonalmatching pursuit, sparse signal reconstruction |
|