A combined scheme in reversible digital image watermarking

One of the reversible watermarking schemes is prediction based reversible watermarking. Sorting technique is a fundamental step in reversible watermarking. If we use effective sorting techniques, better results can be achieved. On the other hand efficient prediction schemes produce smaller prediction errors and embedding in the smaller prediction errors leads to higher quality and capacity. In this paper, a hybrid approach is used. An effective sorting and prediction techniques are used together. Using the artificial neural networks instead of averaging results to predict more accurate and simultaneously using an efficient criterion for sorting instead of the local variance results to an increase in capacity and the visual quality of the watermarked image. The results of this combined approach show that the proposed method is superior to each of the main methods.