摘要
结合人类视觉系统特性及离散余弦变换理论,提出了一种基于原始视频的稳健性盲水印算法。该算法将置乱后的二值图像水印嵌入到原始视频帧绿色分量的离散余弦变换低频系数中,根据视频内容自适应地选择嵌入深度,较好地实现了水印在宿主视频中的嵌入及盲提取。实验结果表明,该视频水印算法在保证不可知性的同时对诸如Gaussian、帧剪切、Salt等攻击均具有良好的稳健性,即使在攻击强度较大的情况下仍然能够提取出较清晰的水印信号。
A robust blind video watermarking algorithm based on original video is proposed combined the HVS and the DCT theory. The algorithm embeded the binary image watermark after scrambling in the DCT low-frequency coefficients of the original video frame＇s green component, and adaptive selection of the embed-depth according to video content. The watermark embedded in the host video and the blind extraction are achieved. The experimental results show that the algorithm has better imperceptibility, and the algorithm are robust against attacks such as the Gaussian, frame shear, Salt, etc, even under the strong attack, it can still be able to extract a clearer definition of watermark signal.
出处
《电视技术》
北大核心
2011年第9期30-32,35,共4页
Video Engineering
基金
江西省教育厅科学技术研究项目(赣教技字[2007]207号)
