A visually secure image encryption scheme based on 2D compressive sensing and integer wavelet transform embedding

Abstract

Using a certain image encryption algorithm to encrypt plain image into visually meaningless noise image is a common way to protect the security of image data. Nevertheless, meaningless noisy image will cause more attention during the transmission process and thus results in lots of attacks. Here, we propose a visually secure image encryption scheme based on two-dimensional compressive sensing (2DCS) and integer wavelet transform (IWT) embedding. During the encryption process, the confusion and diffusion of plain image are realized through random scrambling and grayscale transformation, then a pre-encrypted image is obtained, and then the pre-encrypted image is re-encrypted into a secret image via 2DCS. During the embedding process, the carrier image is converted into a coefficient matrix through IWT, then embedding technology is used to embed the secret image into the coefficient matrix to obtain the modified coefficient matrix, and then it is transformed into the spatial domain through inverse IWT to achieve the final meaningful cipher image. The application of 2DCS can ensure data security and increase the speed of the decryption algorithm, and the IWT embedding can achieve visual security without loss of information. The effectiveness, data security, visual security and real-time performance of the scheme have been tested in simulation results and comprehensive performance analysis.