Reversible Data Hiding Based on Improved Block Selection Strategy and Pixel Value Ordering

Block selection is an effective strategy to improve the performance of pixel value ordering (PVO) based reversible data hiding since image blocks with higher smoothness are first employed to embed secret data, which means a better prediction accuracy could be achieved. However, in current solutions, the definition of smoothness cannot always reflect the true performance of prediction since the calculation only involves limited pixels. Even though more pixels are considered in the state-of-the art, the precision of prediction is still seriously affected by the fluctuating pixels. In this paper, we proposed an improved calculation method for smoothness, which takes enough pixels into account while eliminates the fluctuating pixels simultaneously. Experimental results show that our proposed scheme greatly improves the accuracy of prediction and achieves a better imperceptibility compared with the state-of-the art when the secret bits are not fully embedded.