Research on Video Watermarking Scheme Based on Chaos, DWT and JND Model

The paper presents a novel and robust video watermarking scheme for copyright protection based on chaos, DWT and JND model of HVS. Firstly, in order to ensure security of watermarking and reduce the quantity of computation in the embedding process, we adopt a technique called chaotic selection to select the embedding frames from the video. To every selected frame, we transform it by DWT and detect the variation status of the coefficient in low frequency domain. Because HVS isn't too sensitive to the motive things, we choose those coefficients whose variations are large as the embedding coefficients. Finally, the watermarking signals are embedded and detected according to the JND model. The experimental results show that the proposed watermarking algorithm is robust to additive noise, MPEG compression, frame deleting and so on.