Reversible Video Watermarking Using Motion Estimation and Prediction Error Expansion

In this paper, a novel reversible video watermarking scheme using motion estimation and prediction error expansion is presented, which can embed a large amount of secret data into videos with imperceptible modification. Different with other reversible watermarking schemes based on prediction error expansion, motion estimation is employed to explore the relationship of neighboring frames and work out prediction errors, which significantly sharpen the distribution of prediction errors. Then histogram modification is used to expand the prediction errors, wherein a prediction error is changed by 1 or left unchanged for embedding one secret bit. Due to the slight modification of pixels, high video quality is preserved. In the phases of motion estimation and histogram modification, a little side information used for extracting and recovering are generated, which will be combined with the secret bits and embedded into the video. Experimental results demonstrate that the proposed scheme provides a much larger watermark capacity and a better quality of watermarked video than those of as reported in other state-of-the-art literature.