Adaptive reversible data hiding method based on error energy control and histogram shifting

Tai et al. proposed a reversible data embedding method based on histogram shifting and achieved an excellent embedding performance. Their method shifts the absolute difference of two consecutive pixels for data embedment, and employs an embedding level to control the payload. However, the shifting of absolute difference reduces the number of embeddable spaces and results in a reduction in payload. Instead of shifting the absolute differences, this paper proposes an adaptive method to increase the number of embeddable spaces by referencing a dual binary tree. We also adopt a better predictor and employ an error energy estimator to reduce the number of non-embeddable prediction errors. The experimental results reveal that the proposed method significantly improves the image quality and payload of Tai et al.'s works, especially at low embedding level.