A multi-party secure encryption-sharing hybrid scheme for image data base on compressed sensing

Abstract

A multi-party secure encryption-sharing hybrid scheme for image data base on compressed sensing (CS) is proposed. Traditional image compression and encryption schemes only transmit a single secret image in the network, which is easy to lose or be deceived by a single receiver. In order to solve this problem, the secret sharing is introduced to share and transmit the secret image. The plain image is sparsely processed by wavelet transform, zigzag confusion reduces the correlation between pixels, and the measurement matrix generated by the CS and chaotic system is applied to the sparse image to obtain the secret image. Then the secret image is embedded into the carrier image used least significant bit to form a cipher image, which is also a visually secure image. Sharing divides the cipher image into N parts and transmits them to N recipients for secure transmission, and can be recovered only when r(r≪N) recipients choose to recover the image. Experimental results show that the proposed scheme solves the problem of receiver fraud and network transmission insecurity on the basis of ensuring good recovery quality. In addition, the combination of CS and the semi-tensor product reduces the size of the measurement matrix and reduces the computational complexity. The scheme achieves the effect of hiding and encryption, which can resist steganalysis.