一种小波域内的鲁棒脆弱图像水印技术

目前的水印算法有鲁棒水印和脆弱水印两类，分别可以用于图像的版权保护和篡改检测，但很少有算法能够结合两者，同时实现这两种功能。为了解决这一问题，该文提出一种鲁棒-脆弱水印算法：首先在小波变换图像的逼近子带中，利用系数间的关系构造特征水印；然后根据小波零树思想，结合主分量分析，采用抖动调制方法嵌人到重要小波树的第一主分量中；水印检测时，将提取的特征水印与嵌入水印进行比较，不仅可以确定图像的版权，还可以证明内容的完整性。实验结果表明，此算法能够抵抗滤波、压缩、叠加噪声等处理，同时还可以很好的定位图像的篡改，是一种很有效的水印算法。

A Robust-Fragile Image Watermarking Technique in DWT Domain

At present,watermarking algorithms can be classed into two categories: robust watermarking and fragile watermarking,which can be respectively used to copyright protection and tamper detection.However,hardly any algorithm can incorporate them to achieve both functions.Therefore,a robust-fragile watermarking algorithm is presented,which exploits the correlation of coefficients in the approximate subband of the wavelet transformed image to generate the feature watermark,then exploiting the idea of wavelet zerotree and PCA,embeds it into the significant wavelet trees using DM(Dither Modulation).Comparing the extracted feature watermark and embedded watermark can identify the image's authority and detect the content integrity.Experiments show,the proposed algorithm can resist filtering,compression,adding noise,etc.,while it can locate the image tamper.In a word,it is an effective algorithm.