针对当前视频隐写算法对视频质量和码率影响较大等问题,结合H.264编码标准的宏块分割特性,提出一种基于灰色关联分析的视频隐写算法.该算法对原始载体进行灰色关联度计算,判断其是否为非平滑块,再对帧做离散余弦变换(DCT)变换,根据H.264/AVC编码的宏块分割特性得出的分块大小选择合适的嵌入量,在DCT块的低频区域进行嵌入.实验结果表明,进行秘密信息嵌入后,对视频序列的影响较小,亮度分量的峰值信噪比(PSNR)值平均下降约1.323dB,隐写分析平均检测率为35.73%,算法的平均嵌入容量为417b/帧,对码率影响仅在3%以下,攻击后提取秘密消息的相似度(SIM)值在0.79以上.因此,该算法对视频质量和码率影响较小,并具有抗噪声、滤波攻击,隐写容量大等优点.
Concerning the problem that the existing video steganographic algorithms have large impact on the video quality and the bit rate,a new algorithm based on grey relational analysis combining with the partition modes of H.264/AVC was proposed.The algorithm computed the grey relevancy of blocks and judged it that if they had texture features.Then it did discrete cosine transform(DCT)on the current frame.Finally the algorithm is chosen the proper capacity in terms of the partition mode,and embedded information in low frequency factors of DCT.Experimental results showed that the average decrease of peak signal to noise ratio(PSNR)value was about 1.323 dB.The detection rate was under 35.73%and the average embedding capacity was much larger than 417b/frame.Increment of bitrate was bounded to 3%.The similarity(SIM)values were bigger than 0.79.The experimental result indicates that the proposed algorithm has little impact on the video quality and bit rate,and also has the advantages of antinoise,anti-filtering and high capacity in information hiding.