Abstract: Aiming at the requirements of the copyright information required high capacity and the strong robustness based on MPEG-4, a method of video watermarking which is based on DCT domain in MPEG-4 is proposed. The algorithm synthesizes wavelet transforms, chaos theory and spread frequency technology to do the watermarking pretreatment, and chooses the watermarking embedded domain according to the texture characters of video images. The algorithm, which used for reference of the idea of LSB algorithm, embeds the watermark information into DC component in DCT domain of I-VOP. The experimental results show that the watermark has good invisibility and robustness.

Key words: video watermarking, MPEG-4, DCT, copyright information, robustness