Reversible data hiding of high payload using local edge sensing prediction

This paper tries to analyse a new framework in lossless data hiding research. Its key is how adaptively to get better difference image architectures for given applications. A unique sampled pattern is introduced and described in term of high-similar interpolation image. The proposed method updates a type of conventional algorithm, smoothly accesses to current reconstruction implementation, and outperforms solo weighing approach, particularly in the perceptual quality. Combinational weight factor can be adaptively adjusted by estimating interpolation errors around neighbor pixels. Seeking higher peak value in the difference-image is also our concerns. As an entire algorithm, reversible data hiding based on histogram-difference-shift is also reported. Simulations results demonstrate and verify that our approach is much effective than some recent methods with good generalization performance.