小波域分形编码数字水印的研究

提出了一种新颖的基于小波域分形编码的数字水印技术.分形图像编码方法利用图像的局部自相似性构造分块迭代函数系统,通过保存放射变换参数实现对图像的编码.利用分形变换的编码解码过程中的某些不变参数,可以实现数字水印信息的嵌入和提取.但是空间域的分形编码会使得恢复图像产生明显的块效应现象,嵌入水印后的图像质量损失大.因此一种新的基于小波域细节子图的分形编码方法被提出和研究.利用小波域细节子图的自相似性构建迭代函数系统,实现对细节子图系数的编码.而后基于小波域分形编码,构造了一种新颖的水印嵌入提取算法.算法选取适当的小波细节子图,分块构造匹配池,根据嵌入的水印信号选择最佳匹配块实现水印信息的嵌入和提取.最后通过实验对水印算法受图像处理攻击时的鲁棒性作了分析和比较,嵌入水印的图像在遭受一般图像处理攻击,如滤波、有损压缩、噪声及几何攻击时水印均可读出.构造的基于小波域分形编码的数字水印技术得到了较满意的实验结果.

Digital Watermarking based on Fractal Coding in Wavelet Domain

Digital watermarking technology has attracted a great deal of attention in recent years,because watermarking is good way to protect the intellectual property rights of digital media.The information carried by the watermark can be accessed by using a detection algorithm with the secret key provided.An important property of a watermarking is its robustness with respect to image distortions.This means that the watermark should be readable from images that undergo common image processing operations,such as filtering,lossy compression,noise adding,histogram manipulation,and various geometrical transformations.The fractal theory has proved to be suitable in the field of image coding and watermarking.In general terms,a fractal coder exploits the spatial redundancy within the images by establishing a relationship among its similar parts.It is assumed that many objects can be closely approximated by self-similarity objects that might be generated by use of Iterated Function System(IFS) transformations.According this assumption,the IFS can be seen as a relationship between the whole image and its parts.Thus the main problem is how to find a set of transformations to describe the relationship,in other words,how to define the IFS.Employing the invariant parameters of the fractal transformation,the watermark embedded into the original image can be extracted form the retrieved image.But mosaic effect and bad image quality can be brought by fractal transformation in spatial domain. A novel digital image watermarking method based on fractal coding-decoding and wavelet transformation is presented in this paper.A fractal coding-decoding algorithm in the coefficients of detail images in wavelet domain is presented and studied.On the basis of it,a watermarking scheme is proposed.The watermarking method uses the fractal code which is generated by calculating an Iterated Function System from the detail image.Employing the invariance of the Iterated Function System code,the mark can be hidden and retrieved.In order to greatly improve the robustness and image quality,certain resolution level coefficients in wavelet domain is coded and watermark is embedded into it.The algorithm adds artificial and visually invisible local similarity into the detail image in wavelet domain.This is done by substitution of the range block with a new Matching Block in stead of the original best one.In other words,the watermark is embedded altering the original collage map with a new one which is controlled by the watermark.The proposed watermark embedding consists of a coding-decoding process and the mark retrieving will be performed as a fractal coding process too.Experiments have been performed in order to measure the robustness of the scheme against image processing attack.The watermark is readable from images that undergo common image processing operations,such as filtering,lossy compression,noise adding,histogram manipulation,and various geometrical transformations.In each case,very promising results have been obtained.