Efficient reversible data hiding in encrypted images based on multi-stage integer wavelet transform

TLDR: Experimental results demonstrate that the proposed algorithm significantly improves the embedding capacity and the reconstructed image quality compared with some existing state-of-the-art schemes.

Abstract: Recently, reversible data hiding in encrypted images has drawn increasing interest. We theoretically analyze the feasibility of reversibly hiding secret data into encrypted images using integer transform and propose an efficient reversible data hiding scheme in encrypted images based on integer wavelet transform (IWT) in this paper. The cover image is first encrypted. Then the encrypted image is decomposed with IWT to get coefficients of four sub-bands. Location maps are generated according to the most significant bit-planes of high frequency coefficients. The location maps are further lossless compressed and embedded with secret data into the most significant bits of high frequency wavelet coefficients by bits replacement. Ratio correction (RC) is applied in the pre-processing phase to make sure the efficiency and validity of the proposed scheme. To achieve higher capacity, multi-stage IWT and multi-level embedding are performed. In addition, we then propose an adaptive correction (AC) algorithm which allows the scheme to accommodate more payload. On the receiving side, a recipient can correctly extract embedded secret message with data hiding key. The original image can be perfectly recovered with both the data hiding key and the encryption key. Experimental results demonstrate that the proposed algorithm significantly improves the embedding capacity and the reconstructed image quality compared with some existing state-of-the-art schemes.