A visually secure image encryption scheme based on parallel compressive sensing

It is generally recognized that encrypting an original image into meaningless cipher image is an ideal method to protect image information. However, during transmission, the meaningless cipher image would draw attention and thus attract attacks. Recently, compressive sensing (CS) and carrier images have been utilized by Chai et al. to construct a novel image encryption scheme with visual security. However, in this scheme, some extra transmission is required for possible decryption besides the encrypted image. Moreover, the imperceptibility of the cipher image can be further improved and the recovered image quality would be severely degraded if unsuitable carrier images are selected. In this paper, we design a visually secure encryption scheme by using the parallel compressive sensing (PCS) counter mode and embedding technique. In order to achieve higher security level, Logistic-Tent system and 3-D Cat map are introduced to construct the measurement matrices and to disturb the order of the embedded information, respectively. Furthermore, experimental results demonstrate that the cipher image exhibits superior imperceptibility and the recovered image possesses more satisfactory quality, which is independent of the carrier image.