摘要
小波变换拥有良好的空间——频率局部化特性和与人眼视觉特性相适应的变换机制,但不具有平移、旋转和缩放不变性.采用该类方法的数字图像水印抗噪声攻击和压缩攻击有一定鲁棒性,但抗几何攻击的能力较差.为此提出一种抗几何攻击的小波域数字图像水印算法.算法在水印嵌入时,先对载体图像进行小波分解并对低频子图进行分块,然后对各个子块根据量化规则重复嵌入相应的1bit水印信息;增加低频子图的纹理区的量化步长以提高纹理区噪声掩蔽性.实验结果表明,该水印算法容量较大,提取过程不需要原始图像;有很好不可见性;能抵抗常见几何攻击,对于高斯噪声和JPEG压缩等其它攻击也具有较强的鲁棒性.
DWT having space-frequency localization characteristics and with the transformation mechanism adapted the human visual system,but not having the translation,rotation and scaling invariance.Using the DWT method of digital image watermarking in the resistance to noise attacks and compression attacks,have some robustness,but are less able to resist geometric attacks.This paper presents a geometric attack against digital image watermarking algorithm in wavelet domain.Watermark embedding algorithm,the first wavelet decomposition of the vector image as well as low-frequency sub-block graph,then quantifying the various sub-blocks according to the rules embedded in the corresponding 1bit duplicate watermark information;increase the low-frequency sub-images of the texture area quantization step in order to improve the texture area of noise masking.The experimental results show that the watermarking capacity of large,the extraction process does not require the original image;have a very good invisibility;resistance to common geometric attacks,for Gaussian noise and JPEG compression and other attacks also has strong robustness.
出处
《江西理工大学学报》
CAS
2010年第3期51-55,共5页
Journal of Jiangxi University of Science and Technology
基金
江西省教育厅教改资助项目(JXJG-09-6-8)
关键词
数字图像水印
小波变换
几何攻击
digital image watermarking
wavelet transform
geometric attacks