A Robust Image Blind Watermarking Algorithm Based on Adaptive Quantization Step in DWT
A Robust Image Blind Watermarking Algorithm Based on Adaptive Quantization Step in DWT
复制标题
DOI：
--
复制
发表时间：
2006
影响因子：
--
通讯作者：
Yang Xiao-ping
中科院分区：
文献类型：
--
作者：
Yang Xiao-ping
提出了一种新的基于自适应量化步长的数字水印算法，该算法从LL子带中选取具有感知意义的系数作为水印的嵌入域，采用最小量化步长，保证了水印的不可见性和鲁棒性，对选定的系数进行量化，通过自适应地增加量化步长，将二值水印交替地嵌入到不同的量化层次上，作为一个特殊的例子，该算法利用小波变换在空间域的局部化特性，采用基于块的舍入方法来达到精确嵌入的目的，在水印检测算法上，为了提高系统的鲁棒性，提出了一种新的差分检测算法，实验结果表明，该算法具有以下特点：大的信息容量，不可见性和对攻击的鲁棒性，例如一些常见的图像处理操作，有损压缩，裁剪和一些几何变换等。
A new image blind watermarking algorithm based on adaptive quantization step in discrete wavelet transform(DWT) is proposed,which can be used in copyright authentication.In the proposed algorithm,the perceptually significant coefficients are selected from LL subband to be the embedding domain,and a minimum quantization step,which guarantees the invisibility and robustness,is used to quantify the selected coefficients.The binary watermarks are alternately embedded on the different quantified levels by adaptively increasing the quantization step.As a special example,the least 5 bits of the selected coefficients are utilized to express the watermark.Considering the localization character of wavelet transform in spatial domain,a block-based rounding method is adopted to achieve the purpose of precise embedding.As to the detecting algorithm,a novel difference detecting algorithm is created to improve the robustness of system.The experimental results show that the proposed algorithm has the following characters: large information capacity,invisibility and robustness against attacks such as some common image processing operations,lossy compression,cropping and some geometric transformations etc.