摘要
对原始图像进行3级离散小波变换,对水印先进行金字塔变换,然后对金字塔2级分解的低频系数,和将两次金字塔分解的低频进行扩频后用原始水印图像减去低频系数得到的数进行位平面分解。将位平面分解以后的低4位和高4位分别嵌入到原始图像每一层小波分解后的水平和垂直分量当中,通过逆变换提取出嵌入的数字水印。这种算法保证了水印不可见性,同时对于一般的攻击,例如滤波、噪声、剪切、旋转等具有较强的鲁棒性。
For the original image, it is decomposed by a three-level discrete wavelet transformation(DWT). For the watermarking, firstly it is decomposed by pyramid transformation, then to the low-frequency coefficients of the pyramid decomposition and the numbers that the original image minused the two low frequency coefficients which carried on the wide frequency, they are decomposed by bit-plane decomposition. After that the high and low four bits of the bit-plane decomposition are embedded in each layer in the horizontal and vertical components of wavelet decomposition, through the inverse transformation, extracts the embedded digital watermarking. This algorithm guarantees the watermark’s invisibility, and to the general attacks, such as filtering, noise, shear, rotation, it is quite robust.
出处
《现代计算机》
2010年第6期47-50,共4页
Modern Computer
关键词
数字水印
金字塔变换
位平面
小波变换
Digital Watermarking
Pyramid Transformation
Bit Plane
Wavelet Transformation