An efficient visually meaningful image compression and encryption scheme based on compressive sensing and dynamic LSB embedding

In this paper, an efficient visually meaningful image compression and encryption (VMICE) scheme is proposed by combining compressive sensing (CS) and Least Significant Bit (LSB) embedding. First, the original image (I-orig) is compressed and encrypted into a secret image (I-sec) by CS and Zigzag confusion. Next, dynamic LSB embedding is utilized to randomly embed I-sec into a separate carrier image (I-car) to create the final visually meaningful (VM) cipher image (I-ciph), which is the same size as I-orig. To generate the measurement matrix for CS and determine the embedding position of I-sec, in I-car, a 3-D Cat map is employed. The utilization of this less-chaotic system makes our algorithm easy to implement. Moreover, the initial values of Cat map are computed using I-orig, making our image compression and encryption (ICE) algorithm robust to known-plaintext and chosen-plaintext attacks. Additionally, the adoption of LSB embedding allows for I-car to be selected independently of I-orig. Simulation results and performance analyses are presented to illustrate the effectiveness and efficiency of the proposed cryptosystem.