New sensing approach for compressive sensing using sparsity domain

Compressive sensing is a new technique to acquire sparse signals at a sub-Nyquist rate. This acquisition is done via a sensing step, which has an important impact on the CS performance accuracy. In this paper, a robust new sensing approach is proposed. The idea is to take in consideration the domain of sparsity in the sensing step. Extensive simulations on both synthesized signals and real images are carried out to show the power of the proposed sensing approach compared to the conventional one using different sensing matrices and different reconstruction algorithms.