Adaptive Embedding Combining LBE and IBBE for High-Capacity Reversible Data Hiding in Encrypted Images

Zhifeng Fu,Xiuli Chai,Zongwei Tang,Xin He,Zhihua Gan,Gongyao Cao
DOI: https://doi.org/10.1016/j.sigpro.2023.109299
IF: 4.729
2024-01-01
Signal Processing
Abstract:Reversible data hiding in encrypted image (RDHEI) has received wide attention as an effective information protection technique, and most of the existing RDHEI schemes introduce pixel prediction technique to improve embedding performance. However, the high embedding capacity (EC) requirement is still not achieved because the redundancy of carrier medium is not fully utilized. Thus, a novel RDHEI based on pixel prediction and adaptively combining "L"-shaped block embedding (LBE) and improved binary-block embedding (IBBE) is presented, which maximizes the use of Laplacian-like distribution of prediction-errors to achieve higher EC. Specifically, the original image is first predicted to get the prediction-error image and then eight binary prediction-error bit-planes (PEBPs) are generated. Subsequently, the proposed LBE and IBBE are used to calculate the net EC to encode each PEBP. Also, a new bit-plane selection encryption is designed to ensure that the intra-block correlations of the IBBE-based PEBPs remain unchanged. Eventually, the data hider adaptively completes data embedding according to the PEBP coding. Moreover, the 4-D hyperchaotic system is exploited to encrypt the carrier image and the secret data to further enhance the security. Extensive experimental results verify that our strategy achieves more outstanding performance compared with some state-of-the-art schemes.
What problem does this paper attempt to address?