Hiding cipher-images generated by 2-D compressive sensing with a multi-embedding strategy

In terms of 2-D compressive sensing (CS), multi-embedding strategy and chaotic systems, a novel color image encryption scheme to generate visually meaningful cipher image is proposed in this paper. Firstly, red, green and blue components of color image are compressed to obtain measurement value matrices by virtue of 2-D CS, respectively. Next, the assembled measurement value matrix is shuffled in bit and pixel level and diffused to get a compressed cipher image. Subsequently, two embedding methods are presented to get visually meaningful color or grayscale cipher image by embedding cipher image into color or grayscale carrier image. Compared with volume of plain image, that of carrier image and final cipher image has reduced more than 50%. Moreover, a 4-D hyperchaotic system is applied to generate measurement matrices and control pixel-level confusion, Logistic system is utilized to control bit-level confusion and produce key matrix combined with Local Binary Pattern (LBP), and their initial values are computed by SHA 256 hash value of plain image, which enhances the relationship between plain image and encryption process. Embedding hash value into carrier image prevents extra transmission and storage. Simulation results and security analyses demonstrate the effectiveness and security of the proposed algorithm.

Duke Scholars

Author Yiran Chen Electrical and Computer Engineering