Reversible video watermarking using motion-compensated frame interpolation error expansion

In this study, a reversible video watermarking method based on motion-compensated frame interpolation error expansion is developed. Interframe correlation is exploited more efficiently as a result of using motion-compensated frame interpolation error instead of motion-compensated prediction error that is used in the current reversible video watermarking methods. For this reason, on the one hand the proposed method allows high-capacity data insertion to video; on the other hand, it causes small distortion in the original video. In the proposed method, the original video and watermark can be obtained reversibly from the watermarked video, and the amount of side information in the watermarked video required for watermark decoding and video restoration is extremely low. The method is shown to be superior to the existing methods in terms of capacity and visual video quality by performing computer simulations carried out for various widely used test video sequences