Compare Performance of Recovery Algorithms MP, OMP, L1-Norm in Compressive Sensing for Different Measurement and Sparse Spaces

In this paper, at first, compressive sensing theory involves introducing measurement matrices to dedicate the signal dimension and so sensing cost reduction, and sparse domain to examine the conditions for the possibility of signal recovering, are explained. In addition, three well known recovery algorithms called Matching Pursuit (MP), Orthogonal Matching Pursuit (OMP), and L1-Norm are briefly introduced. Then, the performance of three mentioned recovery algorithms are compared with respect to the mean square error (MSE) and the result images quality. For this purpose, Gaussian and Bernoulli as the measurement matrices are used, where Haar and Fourier as sparse domains are applied.