【作者】 刘金清； 陈超英； 郑孔华； 黄宇升；

【Author】 Liu Jinqing Chen Chaoying Zheng Konghua Huang Yusheng (School of Physics and Optoelectronics Technology,Fujian Normal University, Fuzhou 350007,China)

【机构】 福建师范大学物理与光电信息科技学院；

【摘要】 目前,视频水印算法主要分为两类:基于原始域视频水印算法和基于压缩域水印算法。基于压缩域水印算法多是与具体的视频压缩算法紧密结合,其水印未曾嵌入到视频图像的内容中去,很难承受重编码、二次压缩等处理,其鲁棒性不强。目前H.264压缩标准得到广泛采用,而关于基于H.264的视频水印方案较少,本文提出了一种基于H.264的扩频水印算法。本算法是在视频原始域中嵌入水印,把水印信息嵌入到图像内容中去,让水印与图像经历同样的变化,从而提高其鲁棒性。水印的嵌入是把水印信息经过两次扩频后,嵌入到DCT域直流系数中,而水印的提取则使用线性相关检测原理来进行检测,不需要使用原始载体。从实验结果来看,此算法对H.264压缩及二次压缩具有良好的鲁棒性,嵌入水印后不易被觉察,具有较好的不可感知性。本文具有很高的应用价值。更多还原

【Abstract】 At present, the algorithm of video watermarking mainly is divided into two kinds: based on primitive domain and compression domain. Algorithm of video watermarking based on compression domain mainly relate to video compression algorithm. Because video watermarking is not embedded in image content, it has not strong robustness against the attacks such as re-coding and recompression. In consideration of popularity of H.264 standard and lack of the algorithm of video watermarking based on H.264 standard, so the paper presents a spread spectrum watermarking scheme. In the proposed scheme, the watermarking is embedded in primitive video domain. The advantage of watermarking embed in image content will transform with the image in together, so it has very good performance at Robustness. The watermarking spread spectrum twice is embedded in the DC coefficient in the blocks of discrete cosine transformation (DCT). The watermarking is detected with linear correlation theory, not using primitive video. Experimental results show that the watermarked video is imperceptible and resistant to H.264 compression and recompression.更多还原