Reversible Blind Image Hiding Algorithm Based on Compressive Sensing and Fusion Mechanism

Most current existing image hiding schemes are non-blind, that is, the receiver needs both carrier image and its corresponding container image (i.e., carrier image containing secrets) to extract the hidden secret image. Besides, the existing visual image encryption algorithm suffers from key management and distribution issues. To solve these problems, a reversible blind image hiding algorithm using compressive sensing and fusion mechanism is presented by combining integer wavelet transformation. Besides, a novel mathematical model of key acquisition is established to generate initial keys based on the plain image and a public-key cryptography RSA. Furthermore, a novel nonlinear 4D chaotic map named NewCCM is designed. The test results also explain that the proposed algorithm can reach a good reconstruction effect to the secret plain image without dependence on its corresponding original carrier image, i.e., blind extraction operation is achieved. Moreover, the proposed algorithm performs and displays high security, strong robustness and high reconstruction quality. For example, the information entropy of the obtained cipher image can reach 7.997, the PSNR between carrier image and container image can achieve 33 dB and the NC value can round to 0.999 by a compression ratio set to be half.