A Robust And Transparent Digital Video Watermark Scheme

Due to the rapid development of communication technology, the video transmission is easier. As a result, the copyright protection gets more and more concerns. Traditionally, digital video watermark is widely used for copyright protection. In order to achieve the purpose of copyright protection, the watermark scheme must be robust enough. At the same time, high quality videos become more and more popular. For those videos, the transparency of watermark scheme is very important.Few of current digital video watermark schemes can meet the requirements of both strong robustness and transparency. In this paper, a scheme which can achieve those targets is proposed. In the first step of embed process, watermark information was encoded by split, adding synchronization information, spreading spectrum and scrambling process. Then encoded watermark information was embedded into8*8DCT blocks by changing the coefficients of point (5,2) and (4,3) and the amount change is calculated according to a proposed HVS model. The first step of extract process is calculating DCT coefficients of each8*8block in current frame. The relationship between two points which are changed in embed process defines the original information extracted from current frame. Then the extracted information was decoded by inverse scrambling, inverse spread spectrum, calculating the correct rate of synchronization information and record information. At last, vote detection was applied to detect the final watermark information. The experimental result shows this watermark was transparency and can resist resolution change, format change, bit rate change and frame rate change attacks.