基于混沌和DWT的水印算法

针对数字图像DWT系数的分布特点,分别对在不同频段系数中嵌入和提取水印进行了研究与实验.选用了有意义的灰度图像作为水印,采用基于混沌序列的加密方法对水印进行处理,在水印嵌入域上,选择了具有与人眼视觉特性十分相符的离散小波变换,通过将原始图像进行多级小波分解,在中间某一级或几级的小波系数中嵌入水印信息,从而加大了水印嵌入的灵活性.实验结果表明,在DWT低频域嵌入水印,对算法的透明性影响较大;在中高频域嵌入水印,对透明性的影响相对小一些,并且可加入的水印容量相对较大,但对低通类图像处理鲁棒性较弱.因此可以将各种方案相互融合、相互补充以均衡抵抗各种不同类别的攻击,并根据实际需求进行方案的选择.

Watermarking Algorithm Based on Chaos and DWT

In view of the distributing feature of DWT coefficients,lots of researches and experiments on the watermark embedding methods in different frequency coefficients are made.These two algorithms,in which the gray-level image with practical meaning is selected as the watermark,are encrypted by the method based on Chaos Sequence.In watermark embedding domain,the Discrete Wavelet Transform which has a similar human visual system characteristic is employed.Watermark is embedded into certain one or several sets of Wavelet Coefficients to improve the watermark embedding agility.The experimental result shows that the method of algorithm embedding the watermark in DWT low frequency field has larger effect on the invisibility of algorithm.Comparatively,the method of algorithm embedding the watermark in DWT middle-high frequency field has less effect on the invisibility of algorithm,and it can embed large information capacity.However,it can't withstand the attack of low-pass image processing.Therefore various schemes should be made up in order to resist all kinds of attacks and choose right schemes according to the real need.