A visually meaningful double-image encryption scheme using 2D compressive sensing and multi-rule DNA encoding

Abstract

A visually meaningful double-image encryption scheme using 2D compressive sensing and multi-rule DNA encoding is presented. First, scrambling, diffusing and 2D compressive sensing are performed on the two plain images, and two privacy images are obtained, respectively. Then, the two privacy images are re-encrypted using DNA encoding theory to obtain two secret images. Finally, integer wavelet transform (IWT) is performed on the carrier image to obtain the wavelet coefficients, then the two secret images are embedded into the wavelet coefficients and 2k correction is performed, and the obtained result is processed by inverse IWT to obtain a visually meaningful encrypted image. DNA encoding rules selected for the pixel values of different positions in the two privacy images, and DNA operations performed between the two privacy images and the key streams at different positions are controlled by the chaotic system. The application of 2D compressive sensing reduces the amount of data, thus increasing the encryption capacity of the system. The introduction of DNA encoding theory and the double-image embedding process increases the security of the system. The simulation results demonstrate the feasibility of the scheme, and it has high data security and visual security.

Verifiable visually meaningful image encryption algorithm based on compressive sensing and (t, n)-threshold secret sharing

“…Let CR be the compression ratio, t be the threshold of threshold secret sharing, and B be the block size of BCS. For plain images of size Ń N, the algorithm in [37] has no compression stage, the algorithm in [38] uses parallel CS, the algorithm in [8] uses BCS, and the algorithm in [11] uses 2D CS. The algorithms in [8,11,16,39] do not use threshold secret sharing, and the algorithms in [37][38][39] do not belong to visually meaningful image encryption.…”

“…For plain images of size Ń N, the algorithm in [37] has no compression stage, the algorithm in [38] uses parallel CS, the algorithm in [8] uses BCS, and the algorithm in [11] uses 2D CS. The algorithms in [8,11,16,39] do not use threshold secret sharing, and the algorithms in [37][38][39] do not belong to visually meaningful image encryption. From the perspective of the quality of the reconstructed image, the PSNR value of the reconstruction quality of our algorithm can reach 35.76dB.…”

“…Step8: Apply the RSA algorithm in section 2.4 by using equation (11) to digitally sign the hash values…”

“…Chaotic systems [1,2] were widely used in image encryption due to their characteristics of non-periodicity, pseudo-randomness and initial state sensitivity. In addition, in order to improve the security of the algorithm, researchers have also introduced various technologies to design image encryption algorithms, including DNA encoding [3,4], Boolean networks [5], optical transformations [6,7], singular value decomposition [8], compressive sensing (CS) [9][10][11], and quantum computing [12].…”

A new 2D-HLSM-CECP hyperchaotic map and its application on image encryption using fractal-based permutation and cross coupled diffusion

“…Diffusion is to adjust the pixel value of the image, cover the statistical properties of plain-image into the cipher-image, and implement the even distribution of pixels. A wealth of image encryption algorithms (IEAs) has recently been designed based on numerous technologies, such as substitution box [8][9][10][11], Josephus traversal [12,13], hash functions [14][15][16][17][18], DNA sequence operation [19][20][21], fractal theory [22], compression sensing [23,24], deep learning [25,26] and chaos theory [27][28][29][30][31][32]. Among those different tools, the IEA using chaos theory has been proved their enormous achievement due to the suitable properties of chaotic system (CS) for cryptography, such as the pseudo-randomness of dynamic systems, the non-periodicity of motion trajectories and the sensitivity to initial conditions.…”