Compressive sensing-based image encryption with optimized sensing matrix

Abstract

The dimentional reduction of the signal can be obtained by projecting the signal into a sensing matrix, the technique is known as Compressive Sensing (CS). The CS also provides a mechanism for data security (encryption) because the signal can only be reconstructed if the dictionary and the sensing matrix are known. This technique is better compared to the conventional technique because the compression and encryption is done simultaneously through the signal projection into the sensing matrix. Moreover the signal projection into the sensing matrix is relatively simple and doesn't need heavy computational load so it will be very effective to implement in a portable device. We proposed a scheme of simultaneous image compression-encryption based on CS by using a novel method of optimized sensing matrix. The simulation results showed that by using optimized sensing matrix improved the quality of reconstructed image compared with random sensing matrix that is usually used in CS-based encryption.