Controllable high-capacity separable data hiding in encrypted images by compressive sensing and data pretreatment

In this paper, a separable data hiding algorithm in encrypted images with controllable and high-capacity is proposed. Firstly, the original image is decomposed into important coefficients and unimportant coefficients by discrete wavelet transform (DWT). Secondly, DWT is applied again on unimportant coefficients matrixes and then the obtained coefficients matrixes are compressed using compressive sensing (CS) to empty space for data hiding. All the important coefficients are encrypted using the traditional stream cipher by the content owner. Thirdly, the data hider hides the pretreated data information in the free space. Finally, the encrypted image containing additional data is scrambled to improve the security. The receiver can separably extract the hiding data or/and decrypt the image depending on the keys he owns. Compared with the previous work, there are various advantages in the proposed algorithm, such as the separability between image recovery and data extraction, the controllable and high-capacity for data hiding. Experimental results verify the superiority of the proposed algorithm.