A Video Watermarking Scheme Based on 3D-DWT and Neural Network

In this paper, we propose a novel digital video watermarking scheme based on 3D-DWT and artificial neural network. First, a 3D wavelet transform is performed on each selected video shot, then the watermark is embedded in the LL subband wavelet coefficients of the low-pass frames, therefore the scheme has strong robustness against common video processing attacks. At the same time, in order to guarantee the perceptual invisibility, the coefficient is selected adaptively to embed the watermark and the embedding intensity is adaptively controlled using the statistical characteristics (mean and standard deviation). In addition, embedding and extraction of the watermark are based on the relationship between a wavelet coefficient and its neighbor's. A neural network is given to memorize the relationships among the coefficients in a 3×3 block. The watermark extraction process does not require the original video. Experimental results show that the embedded watermark is robust and invisible.