摘要
提出一种MPEG-4盲视频水印算法:二值水印图像经Arnold变换扩频并与伪随机序列调制后得到的水印信息相同地嵌入到每一视频帧Y分量的DCT低中频系数中,通过水印位来调整这些系数的大小,然后根据它们对嵌入强度的余数范围来提取水印位,该算法简单易实现。实验结果表明:本算法具有很好的不可见性,并对JPEG压缩、涂改、椒盐噪声和剪切等攻击都有很好的鲁棒性。
A bind MPEG-4 video watermarking algorithm is proposed. A binary watermarking image transformed with Arnold scrambling is extended. The extended sequence is modulated with pseudo random sequence and then is embedded into the DCT- Coefficients of the low and middle frequency in every video frame. Simultaneity, adjust these DCT-Coefficients according as the watermark bits. Extract the watermarking with the scope of the remainder which these DCT-Coefficients have been divided by the embedded strength. The algorithm is implemented simply. Experimental results demonstrate that the watermarking is invisible and robust to attacks such as JPEG, altering, salt & pepper noise and cropping.
出处
《计算机与现代化》
2010年第2期39-42,46,共5页
Computer and Modernization