An Adaptive Feedback Based Reversible Watermarking Algorithm using Difference Expansion

In this paper an adaptive feedback based Reversible Watermarking (RW) algorithm using Difference Expansion (DE) is designed for gray-scale still pictures. The algorithm obtains automatically the highest payload size and the corresponding embedding threshold and also determines the best watermarked picture for an input cover picture with respect to the quality metric Peak-Signal-to-Noise Ratio (PSNR). These automatically obtained parameters are non-user-defined like other non-feedback based RW algorithms. The payload that can be embedded into a cover image is discussed, and a feedback mechanism to control this size is designed. Software implementation results using MATLAB 2013b for ten tested images are presented and compared with each other based on their quality metrics. The proposed algorithm is then compared with existing non-feedback based RW algorithms. These results indicate that the proposed algorithm has low timing complexity over other existing non-feedback based RW algorithms which in turn provide higher speed.