Reversible watermarking based on histogram shift and adaptive embedding

This paper presents an adaptive reversible watermarking algorithm, which employs histogram shifting and adaptive embedding to embed data into the image. Histogram shifting is a useful reversible watermarking technique which is large in capacity and low in distortion. With adaptive embedding, our method can adaptively embed 1 bit or more bits into an expandable pixel according to the calculated number of embedded bits. This method, on the one hand, contributes to higher capacity and the embedding rate can be greatly improved. On the other hand, we propose BitNum: the number of embedded bits. Computed with the prediction error of the neighboring pixel, BitNum indicates directly how many bits of information should be embedded into an expandable pixel. Besides, pixel selection is an effective way to improve performance of adaptive embedding. The superiority of the performance of the proposed method is showed by different images and the comparison with other methods experimentally. © 2013 ICIC International.