Robust video watermarking scheme based on 3D DCT

In this paper, a novel video watermarking scheme based on 3D DCT(3-dimensional discrete cosine transform) is proposed. Firstly, the video is segmented into several shots. Then, those frames between two neighbor shot frames are grouped into 3D blocks. The watermark is embedded into those middle frequency coefficients of the 3D transform domain. The watermarked video has good transparency and robustness against several attacks, including mpeg compression, frame dropping, noisy and so on. The experimental results against Wang´s method testify the robustness.