Reversible data hiding in encrypted image using new embedding pattern and multiple judgments

Abstract

This work proposes a data embedding method in encrypted images. The proposed method introduces a new embedding pattern which using the pixel spatial correlation in a block and border pixels on the block sides to achieve a low error rate effectively. In the extraction end, the neighboring blocks are also adopted to predict the current block to calculate the smoothness more accurately. Experimental results show our proposed method performs better than Zhang's method and Hong's method. In the case of the block size is 8ź×ź8, the error rate for the cover image Lena in Zhang's method is 1.25 % and the error rate of Hong's method is 0.34 %, while the error rate in our proposed embedding method reduces to 0.2 %.