基于SIFT和非采样Contourlet变换的鲁棒水印算法

针对常见抗几何攻击水印算法存在的不足,设计一种基于SIFT和非采样Contourlet(NSCT)变换的鲁棒水印算法。算法首先利用尺度不变特征变换在载体图像的尺度空间上提取特征点并形成矩形特征区域,然后水印信息自适应量化嵌入特征区域的NSCT低频变换系数中,最后保存若干个稳定特征点用于几何校正。实验结果表明,该算法不仅具有很好的透明性,同时对常见图像处理攻击和几何攻击均具有很强的鲁棒性。     

一种感兴趣区域二维水印算法的实现

针对图像中用户选定的感兴趣区域ROI(Regions of interest)的保护,提出形状自适应离散小波变换SA-DWT(Shape Adaptive DWT)是实现感兴趣区域二维水印算法的有效工具.水印图像(一幅灰度图像)经压缩后生成可以嵌入的水印序列,同时SA-DWT变换作用于ROI区域,通过选择变换系数的嵌入位置和嵌入准则实现水印信息的嵌入;最后提取水印信号,通过重构恢复水印图像.攻击实验表明该算法可以抵抗常见的图像攻击手段,ROI区域得到了有效保护.     

基于感兴趣区域的图像水印嵌入算法研究

鉴于现有绝大多数图像水印方案将载体图像看作整个场景,并未考虑载体图像自身的结构特征及人们感兴趣的视觉目标区域,本文提出了一种基于感兴趣区(ROI)的小波域图像水印嵌入算法.该算法首先结合人眼视觉感知特性,在小波变换域内利用k-均值聚类提取出感兴趣区域;然后对感兴趣区域进行小波变换,并结合图像局部相关性对小波系数进行自适应的量化调制,将水印信息嵌入到感兴趣区域的低频子带内.仿真实验表明,所提出的水印嵌入算法不仅具有较好的透明性,而且对JPEG压缩、叠加噪声、锐化等常规图像处理操作具有较好的鲁棒性.特别地,该算法能够有效抵御恶意剪切、目标移动、替换背景等攻击,而且可以实现盲检测.     

基于Contourlet域特征点的抗几何攻击水印方法

有效抵抗几何攻击的水印算法是目前水印技术研究的难点之一,很小的几何变形就可能使大多数水印算法失效.对图像Contourlet变换和Harris-Affine算子的特性进行了分析和讨论,在此基础上提出了一种以仿射不变特征点的基于Contourlet的抗几何攻击水印算法.首先,用Contourlet变换提取出多尺度、多方向的纹理信息;再用Harris-Affine检测算子从变换域的中、低频方向子带中提取出仿射不变特征点,结合自适应局部结构的特征尺度确定特征区域,并用U变换对其归一化处理,水印就自适应地嵌入到归一化后的区域中;水印嵌入采纳矢量量化的策略,将水印信息重复嵌入到不同方向子带、多个不相交的局部仿射不变特征区域.检测时,利用特征点实现水印的重同步,采用多数原则来提取水印的位信息,无需求助于原始图像.仿真结果表明,算法不仅具有较好的透明性,而且对常规信号处理和去仿射攻击均具有较好的鲁棒性     

一种改进的抗几何攻击彩色图像水印算法

提出一种基于加速鲁棒性特征算法（Speed-Up Robust Features,SURF）的抗几何攻击彩色图像水印算法。首先,在彩色图像中的红色分量中检测SURF特征点,并且生成SURF特征描述算子。然后,利用二级离散小波变换（Discrete Wave-let Transform,DWT）得到图像蓝色以及绿色分量的低频信息,并对其进行奇异值分解（Singular Value Decomposition,SVD）。最后将经过一级小波变换后的水印图像对应嵌入彩色图像蓝色分量和绿色分量的低频区域之中。实验结果表明,嵌入水印后的图像在经过几何攻击后,其提取的水印与原水印仍具有较高的相似度,该算法具有较好的鲁棒性。     

基于纹理特征定位的数字图像水印算法

提出一种基于图像纹理特征锁定水印嵌入位置的水印算法。利用Canny算子在非采样Contourlet变换最高阶方向子带内筛选图像的纹理特征,将强纹理特征的位置作为水印嵌入的索引点。对非采样Contourlet变换的低频子带进行一级小波分解,在小波细节子带上分割N个圆环(N为水印信息长度)区域,根据索引位置在每个圆环区域内自适应地重复嵌入水印。仿真实验表明,该算法不仅具有较好的透明性,而且对图像处理和几何攻击均具有较好的鲁棒性。     

基于小波变换和神经网络的盲数字水印算法

文中提出了一种基于小波变换和神经网络的盲数字水印算法。算法嵌入的水印信息是一幅二值图像,每位水印信息被连续多次嵌入宿主图像小波分解的低频子带。利用提取含水印图像小波分解的低频子图的特征点坐标集估计水印图像遭受的几何失真,根据估计的几何失真参数对受到几何攻击的水印图像进行校正,最后利用神经网络很好的非线性映射和自适应学习功能实现水印信息的盲提取。实验表明,算法对常见的图像处理和几何攻击具有很好的鲁棒性。     

一种基于图像特征的彩色图像水印算法

为了提高水印算法的抗几何攻击能力,提出一种基于图像特征的彩色图像水印算法。首先利用加速鲁棒性特征SURF(Speed-Up Robust Features)算法提取出图像中稳定的特征点,并为其构造SURF特征描述算子,使提取出的特征点具有尺度不变性;然后根据特征尺度的自适应性获得局部特征区域并对其进行归一化处理;之后对局部特征区域进行斜变换,并获得其中频系数;最后结合奇异值分解SVD(Singular Value Decomposition)算法完成水印的嵌入。实验结果表明,嵌入水印后的图像在经过噪声以及几何攻击后,其提取的水印与原水印仍具有较高的相似度,该算法具有较好的鲁棒性。     

基于QR分解的Contourlet域抗几何攻击水印算法

为了提高数字水印产品的抗几何攻击能力,提出一种基于QR分解的Contourlet域抗几何攻击水印算法。该算法首先对宿主图像Contourlet变换后的低频子带进行分块QR分解,通过轻微修改各子块Q矩阵第一列第二行元素和第三行元素的大小关系嵌入Arnold置乱后的水印信息,并且利用R矩阵对各子块进行相应的视觉失真补偿操作。在水印提取前,利用尺度不变特征变换SIFT特征点具有的旋转、缩放和平移不变性对几何失真含水印图像进行几何校正,恢复水印的同步性,使算法具备抵抗几何攻击的能力。实验结果表明,该算法的不可见性良好,而且对于常规攻击和多种几何攻击都具有较好的鲁棒性。     

新式灰度图像盲检测数字水印算法

将冗余离散小波(RDWT)与矩阵Schur分解结合,提出了一种新式灰度图像盲检测数字水印算法。该算法将载体图像进行二级RDWT变换提取其逼近子图低频部分,并将分块后的每个块低频系数进行Schur分解,通过修改Schur分解后的上三角矩阵对角线上的最大能量元素完成对经Fibonacci变换加密后的水印图像的嵌入,水印的提取过程实现了盲提取。实验结果表明,该水印算法水印嵌入的时间仅为0.97s,平均提取水印时间为0.211s,与类似盲水印算法相比,不可见性的PSNR值提高了14.12%,抵抗高斯噪声攻击、椒盐噪声攻击和剪切几何攻击的能力也分别提高了21.32%、17.31%和27.83%。     

A new SVM-based image watermarking using Gaussian-Hermite moments

Geometric attack is known as one of the most difficult attacks to resist, for it can desynchronize the location of the watermark and hence causes incorrect watermark detection. It is a challenging work to design a robust image watermarking scheme against geometric attacks. Based on the support vector machine (SVM) and Gaussian-Hermite moments (GHMs), we propose a robust image watermarking algorithm in nonsubsampled contourlet transform (NSCT) domain with good visual quality and reasonable resistance toward geometric attacks in this paper. Firstly, the NSCT is performed on original host image, and corresponding low-pass subband is selected for embedding watermark. Then, the selected low-pass subband is divided into small blocks. Finally, the digital watermark is embedded into host image by modulating adaptively the NSCT coefficients in small block. The main steps of digital watermark detecting procedure include: (1) some low-order Gaussian-Hermite moments of training image are computed, which are regarded as the effective feature vectors; (2) the appropriate kernel function is selected for training, and a SVM training model can be obtained; (3) the watermarked image is corrected with the well trained SVM model; (4) the digital watermark is extracted from the corrected watermarked image. Experimental results show that the proposed image watermarking is not only invisible and robust against common image processing operations such as filtering, noise adding, JPEG compression, etc., but also robust against the geometric attacks.     

A robust content based audio watermarking using UDWT and invariant histogram

Desynchronization attack is known as one of the most difficult attacks to resist, for it can desynchronize the location of the watermark and hence causes incorrect watermark detection. It is a challenging work to design a robust audio watermarking scheme against desynchronization attacks. Based on undecimated discrete wavelet transform (UDWT) and invariant histogram, we propose a new content based audio watermarking algorithm with good audible quality and reasonable resistance toward desynchronization attacks in this paper. Firstly, the undecimated discrete wavelet transform (UDWT) is performed on original host audio. Secondly, the invariant histogram is extracted from a selected wavelet coefficients range in the low frequency subband. Then, the bin of histogram is divided into many groups, each group including four consecutive bins. For each group, one watermark bit is embedded by reassigning the number of wavelet coefficients in this group of four bins. Finally, the digital watermark is embedded into the original audio signal in UDWT domain by modifying a small set of wavelet coefficients. Simulation results show that the proposed watermarking scheme is not only inaudible and robust against common signal processing operations such as MP3 compression, noise addition, and low-pass filtering etc, but also robust against the desynchronization attacks such as random cropping, time-scale modification, pitch shifting, and jittering etc.     

Geometrically invariant image watermarking using SVR correction in NSCT domain

Based on the support vector regression (SVR) geometric distortions correction, we propose a robust image watermarking algorithm in nonsubsampled contourlet transform (NSCT) domain with good visual quality and reasonable resistance toward geometric attacks in this paper. Firstly, the NSCT is performed on original host image, and corresponding low-pass subband is selected for embedding watermark. Then, the selected low-pass subband is divided into small blocks. Finally, the digital watermark is embedded into host image by modulating the NSCT coefficients in small blocks. In digital watermark detecting procedure, the SVR geometrical distortions correction is utilized. Experimental results show that the proposed image watermarking is invisible, and robust against common image processing and some geometrical attacks.     

Robust digital watermarking in PDTDFB domain based on least squares support vector machine

Geometric distortion is known as one of the most difficult attacks to resist, for it can desynchronize the location of the watermark and hence causes incorrect watermark detection. It is a challenging work to design a robust image watermarking scheme against geometric distortions. Based on the least squares support vector machine (LS-SVM) geometric distortions correction, we propose a new image watermarking scheme in shiftable complex directional pyramid (PDTDFB) domain with good visual quality and reasonable resistance toward geometric distortions in this paper. Firstly, the PDTDFB decomposition is performed on the original host image. Then, the corresponding lowpass subband is divided into small blocks. Finally, the digital watermark is embedded into host image by modulating the selected lowpass PDTDFB coefficients in small blocks. The main steps of digital watermark detecting procedure include: (1) the PDTDFB decomposition is performed on the test images, and some low-order Gaussian–Hermite moment energy of highpass subbands are computed, which are regarded as the effective feature vectors; (2) the appropriate kernel function is selected for training, and a LS-SVM training model can be obtained; (3) the watermarked image is corrected with the well trained LS-SVM model; and (4) the digital watermark is extracted from the corrected watermarked image. Experimental results show that the proposed image watermarking is not only invisible and robust against common image processing operations such as filtering, noise adding, and JPEG compression etc, but also robust against the geometrical distortions.     

人眼视觉特性与SIFT相结合的视频双水印算法

针对视频水印容易遭受几何攻击以及水印的鲁棒性与透明性的平衡问题,提出一种基于人眼视觉特性与尺度不变特征变换（SIFT）相结合的抗几何攻击视频双水印算法。首先获取视频序列中人眼视觉掩蔽阈值作为水印的最大嵌入强度。其次,将视频帧进行离散小波变换（DWT）,对中高频子带系数提出基于视频运动信息的自适应水印算法;针对低频子带,提出基于小波低频系数统计特性的抗几何攻击视频水印算法。最后,以SIFT作为触发器判断视频帧是否遭受几何攻击,对遭受几何攻击的视频帧利用SIFT的尺度与方向不变性进行校正,并对校正后的视频帧提取水印信号;针对非几何攻击的视频帧,直接利用中高频提取算法。所提算法与实时性视频水印算法——基于小波域直方图的视频水印（VW-HDWT）算法比较,峰值性噪比值（PSNR）提高了7.5%;与基于特征区域的水印算法相比,水印嵌入容量提高约10倍。实验结果表明,在保证水印透明度较好的情况下,所提算法对常规几何攻击具有较强的鲁棒性。     

基于几何校正与非下采样Shearlet变换的图像水印算法

为了解决当前图像水印技术难以抵御几何失真,使其鲁棒性较低与误检率较高的问题,提出了几何校正与非下采样Shearlet变换的图像水印算法.首先,引入Cat映射,对水印信息图像进行置乱;随后,借助非下采样Shearlet变换机制,对载体图像进行处理,获取低通子带和高通子带,并将低通子带分割为尺寸相同的小块;通过修改低通子带的Shearlet系数,建立水印嵌入机制,将水印信息植入到载体图像中,获取水印密文;构建几何失真图像训练样本,基于极谐变换,计算水印图像的极谐变换系数模,充分描述其鲁棒特征;基于模糊支持向量机,预测几何失真参数,对水印图像进行几何校正;最后,再次利用非下采样Shearlet变换处理校正水印图像,获取低通子带小块,设计水印提取方法,复原其水印信息.实验结果显示：与当前图像水印算法相比,所提算法具有更高的不可感知能力与鲁棒性,对于各种几何攻击,所提取技术的复原水印与初始水印的相关系数均要高于0.95.     

信息嵌入位置自适应选择的鲁棒水印算法

为了更好地提高嵌入水印后的图像质量,提出了一种利用遗传算法(GA)优化支持向量回归机(SVR)的鲁棒水印算法。把经Haar小波变换后,图像子带中具有强相似性的数据作为特征向量,用于被遗传算法优化的SVR建立小波系数方向树的模型。通过比较特征向量均方差(MSE)的大小来自适应地确定水印嵌入的位置。水印的嵌入与提取是通过调整模型的预测值与目标值之间的大小来实现的。实验结果表明,所提算法对常见的图像攻击有很强的鲁棒性,而且水印图像在嵌入容量为16384比特的情况下,峰值信噪比可以达到44.15dB。因此能够有效抵抗常见的水印攻击,在嵌入大量信息的情况下,具有很高的透明性。