基于DCT的大容量二值水印嵌入新算法

A novel high capacity binary watermarking algorithm based on DCT

传统的基于DCT变换的二值水印嵌入算法通常是在分块系数矩阵中选择一对中频系数实现1bit水印信息的嵌入,水印嵌入容量有限,为提高载体图像的安全性,提出了一种新的基于DCT的大容量二值水印嵌入算法。本算法首先对二值水印图像进行Arnold置乱加密处理,其次对所选取的载体图像进行分块DCT变换,然后选择4个中频系数进行排序,充分利用两两之间的3个差值关系实现水印的嵌入。若待嵌入水印位为0,通过改变系数值使得差值大于或等于阈值A,若待嵌入水印位为1,通过改变系数值使得差值等于阈值B,最终使得每块选择4个系数值实现了3bit二值水印信息的嵌入,水印嵌入容量大大优于传统方法。仿真实验表明,本算法实现了大容量水印的嵌入与提取且抵抗图像裁剪、图像加噪、JPEG压缩等攻击效果较好。

For embedding in the cover image are mostly binary watermarking algorithms based on DCT,the capacity of traditional methods which select one pair of middle capacity frequency coefficients to embed one bit watermark is limited.In order to increase the safety of cover images,this paper proposes a novel high capacity binary image watermarking algorithm based on DCT.Firstly,the binary watermark is preprocessed to enhance the security by Arnold scrambling.Secondly,the paper appliesblock DCT to the selected cover image,then adopts four middle frequency coefficients to reorder and makes full use of differences among the above coefficients to embed watermarks.If the watermark bit to be embedded is 0,so that by changing the coefficient value to make differences are greater than or equal to the threshold value A,and if the watermark bit to be embedded is 1,so that by changing the coefficient value to make differences are equal to the threshold value B.Finally,three binary watermark bits are embedded in each block by changing four middle frequency coefficients,and watermark embedding capacity is much better than traditional methods.The simulation experiments show that the method achieve embedding and extraction of the large capacity watermark and has strong robustness to cutting,image adding noise,JPEG compression,and other attacks.