Robust wavelet-based watermarking scheme for video copyright protection

A novel blind video watermarking scheme is proposed. Along the temporal axis, separate the host video into nonoverlapping groups of pictures (GOPs) in the method first. To keep transparency and robustness, the binary watermark is superposed on the low frequency wavelet sub-bands of GOP by using a quantization algorithm. In the extraction stage, the original host video is not required. Compared with conventional video watermarking method, the suggested scheme provides better performances for robustness against attacks and visual quality. Experiments show that the suggested method is transparent and robust against some attacks, including temporal attack, filtering, noise addition, compression, translation, brightening, and temporal attacks, etc.