一种改进的伪3D-DCT域的视频零水印算法

为进一步提高视频水印算法的鲁棒性,提出了一种改进的伪三维离散余弦变换(3D-DCT)的视频零水印算法。该算法首先采用帧间欧氏距离法选取关键帧,然后,利用三帧差分法得到关键帧的运动目标,并对关键帧的运动目标图像进行分块,同时对运动目标的几何形心坐标所在块进行伪3D-DCT变换。将变换后的AC系数用于构造视频的特征值序列,并将混沌映射加密的水印嵌入特征值序列中构造零水印。最后将零水印在数据库中进行注册。实验表明,该算法对噪声、滤波攻击,帧剪切、旋转、缩放等攻击具有很强的不可见性和鲁棒性。

An improved video zero-watermarking algorithm in pseudo 3D-DCT domain

In order to improve the robustness of the video watermarking algorithm, we propose an improved video zero-watermarking algorithm in pseudo 3-D discrete cosine transform (3D-DCT) domain. In this scheme, we first select key frames through the inter-frame Euclidean distance algorithm, and then apply the three frames difference algorithm to extract the moving object of the key frame. Afterwards the moving object of the key frame is segmented as blocks, and the pseudo 3D-DCT is performed on the blocks. After the sequences of video eigenvalues are constructed by AC coefficients, we embed watermark signals into the sequences of eigenvalues to structure zero-watermark. Finally, the zero-watermark is registered in a database. Experimental results show that the proposed algorithm is invisible and robust against most attacks, including noise filtering attacks, frame cropping, rotation, scaling and so on.