A multiplicative video watermarking robust to H.264/AVC compression standard

The widespread application of the H.264 video compression standard necessitates proper design of copyright protection algorithms. In this paper, a new blind multiplicative video watermarking algorithm robust to this standard and also other attacks is proposed. With this method, the multiplicative watermark is embedded in the HH subband of the wavelet-transformed video frames. The HH coefficients of the host signal are divided into two patches: One patch is used to host the watermark, while the other patch is remained intact to be exploited at the receiver for estimating the system parameters required for watermark detection. Assuming the wavelet HH coefficients to have the Laplacian distribution, the maximum likelihood (ML) criterion is used for the watermark detection. In this case, the ratio of summation of samples of two patches is used as our decision variable on which the ML criterion is applied. Simulation results show that along with the exhibition of an acceptable robustness against various attacks, the proposed algorithm properly satisfies the invisibility requirements of the watermarking concept with low computational complexity.