Wavelet based sparse image recovery via Orthogonal Matching Pursuit

Compressed sensing is an emerging technique that reconstructs signals and images discarding the Shannon Nyquist theory of reconstruction. This paper demonstrates that if orthogonal matching pursuit is implemented in multi stages, it gives a faster recovery of an image with k^th Sparsity level by taking k ln R measurements for a dimension R. The results of orthogonal matching pursuit are comparable with the least squares method of recovery and the numbers o f measurements are comparable with the previous work on Orthogonal Matching Pursuit. It can be taken as an attractive alternative to other recovery algorithms as in some cases at the cost of certain parameters orthogonal matching pursuit is easily implemented and is faster.