Robust Reversible Data Hiding in Encrypted Image

this paper presents a robust version of reversible data hiding method in encrypted images previously proposed by Zhang [8] and Wien [9]. In the original method, the image is partitioned into non-overlapping blocks, each divided into two neighbor sets. In these sets, one bit is embedded by flipping three LSBs of predefined pixels. The data extraction can be done by examining the block smoothness of neighboring blocks. However, this method is very sensitive to noise and attacks. Because noise and malicious activity usually alter adjacent blocks in the same manner. So, neighboring blocks do not have the same smoothness anymore. In this paper we propose a new method to select non-neighbor sets to decrease the effects of noise and therefore reduce bit error rate in encrypted image. These non-neighbor sets are selected by finding the most similar pairs as performed in the encoding phase of fractal image compression. Experimental results show that the proposed method outperforms previous approaches in the case of Gaussian noise contamination and results lower bit error rate.