A blind watermarking algorithm based on DWT and SVD

This paper presents a new digital image blind watermarking algorithm based on combination of discrete wavelet transform （DWT） and singular value decomposition （SVD）. First of all, we make wavelet decomposition for the original image and divide the acquired low frequency sub-band into blocks. Then we make singular value decomposition for each block and embed the watermark information in the largest singular value by quantitative method. The watermark can be extracted without the original image. The experimental results show that the algorithm has a good imperceptibility and robustness.

Robust Audio Blind Watermarking Algorithm Based on Haar Transform

To solve the problem of low capacity of audio watermarking information and poor robustness of impact resistance,a digital audio watermark algorithm based on wavelet transform is presented in this paper. In this algorithm,the fine and approximate components are obtained when the Haar wavelet base is used to convert each frame of the audio signal to its discrete transform. Then the HAS algorithm is applied to the fine component embedded with watermark to realize encryption process. The original audio carrier signal is not needed in extraction,as long as the signal is embedded after the frame is divided. The first two sections of each frame are implemented in Haar transform respectively to get the fine and approximate component. The watermark images are obtained from the former two fine components of the energy watermark sequence after calculating,comparing and extracting. The simulation results show that there is a certain transparency for the algorithm,a good robustness against the resampling and low pass filtering attack.

Robust blind image watermarking based on interest points

Digital watermarking technology plays an essential role in the work of anti-counterfeiting and traceability.However,image watermarking algorithms are weak against hybrid attacks,especially geometric attacks,such as cropping attacks,rotation attacks,etc.We propose a robust blind image watermarking algorithm that combines stable interest points and deep learning networks to improve the robustness of the watermarking algorithm further.First,to extract more sparse and stable interest points,we use the Superpoint algorithm for generation and design two steps to perform the screening procedure.We first keep the points with the highest possibility in a given region to ensure the sparsity of the points and then filter the robust interest points by hybrid attacks to ensure high stability.The message is embedded in sub-blocks centered on stable interest points using a deep learning-based framework.Different kinds of attacks and simulated noise are added to the adversarial training to guarantee the robustness of embedded blocks.We use the ConvNext network for watermark extraction and determine the division threshold based on the decoded values of the unembedded sub-blocks.Through extensive experimental results,we demonstrate that our proposed algorithm can improve the accuracy of the network in extracting information while ensuring high invisibility between the embedded image and the original cover image.Comparison with previous SOTA work reveals that our algorithm can achieve better visual and numerical results on hybrid and geometric attacks.

A blind watermarking algorithm for 3D mesh models based on vertex curvature

We propose a robust blind watermarking algorithm for three-dimensional(3D)mesh models based on vertex curvature to maintain good robustness and improve visual masking in 3D mesh models.In the embedding process,by using the local window of vertex,the root mean square curvature is calculated for every vertex of the 3D mesh model and an ordered set of fluctuation values is obtained.According to the ordered fluctuation values,the vertices are separated into bins.In each bin the fluctuation values are normalized.Finally,the mean of the root mean square curvature fluctuation values of the vertices in each bin is modulated to embed watermark information.In watermark detection,the algorithm uses a blind watermark extraction technique to extract the watermark information.The experimental results show that the algorithm has a very good performance for visual masking of the embedded model and that it can resist a variety of common attacks such as vertex rearrangement,rotation,translating,uniform scaling,noise,smoothing,quantization,and simplification.

A Blind Watermarking Algorithm Resisting to Geometric Transforms Based on SVD

A new watermarking algorithm resisting to geometric transformation based on singular value decomposition （SVD） in logarithm polar coordinate is proposed. The log-polar mapping （LPM） is used to resist rotation and scaling attacks, and the odd-even quantization algorithm is used to embed watermark so it can be extracted without the original host image. The experiments show that the proposed algorithm not only resists various geomet- ric attacks but also is robust enough to the common signal processing.

Hybrid Successive CFA Image Encryption-Watermarking Algorithm Based on the Quaternionic Wavelet Transform (QWT)

In this paper, we present a new robust hybrid algorithm combining successively chaotic encryption and blind watermarking of images based on the quaternionic wavelet transform (QWT) to ensure the secure transfer of digital data. The calculations of the different evaluation parameters have been performed in order to determine the robustness of our algorithm to certain attacks. The application of this hybrid algorithm on CFA (Color Filter Array) images, allowed us to guarantee the integrity of the digital data and to propose an autonomous transmission system. The results obtained after simulation of this successive hybrid algorithm of chaotic encryption and then blind watermarking are appreciated through the values of the evaluation parameters which are the peak signal-to-noise ratio (PSNR) and the correlation coefficient (CC), and by the visual observation of the extracted watermarks before and after attacks. The values of these parameters show that this successive hybrid algorithm is robust against some conventional attacks.

基于顶点范数三维网格模型抗简化盲水印算法

为提高三维网格模型的数字水印对网格简化攻击的鲁棒性,提出一种基于顶点范数的三维网格模型抗简化盲水印算法。结合网格模型点与面的权重关系计算模型中心,利用中心计算顶点范数,根据水印长度建立顶点范数分组,在嵌入水印时,通过映射函数调整分组内顶点范数均值嵌入水印信息;在检测水印时,计算各集合内均值进行水印信息的盲检测。实验结果表明,该水印算法具有较好的透明性、鲁棒性以及简化攻击抵抗的能力,为三维网格模型的版权保护提供了有效技术支持。

Embedding Color Watermark Image into Color Host Image Based on Contourlet-Domain

A novel image blind watermarking scheme based on Contourlet transform was proposed to embed color watermark image into color host image, which was different from some existing works using the binary or gray-level image as watermark. In the process of embedding watermark, each color component of the color host image was decomposed by Contourlet transform. Then, the low-pass sub-band coeffidents were divided to 4 x 4 non-overlapping blocks, and the eoeffidents in the selected block were quantified for embedding the watermark information. In the process of extraction watermark, the watermark could be extracted from the watermarked image without the help of the original host image or the original watermark image. Experimental results show that the proposed color image scheme has better invisibility and stronger robustness against most attacks such as image compression, filtering, adding noise, cropping, rotation, and scaling.

Wavelet Transform Modulus Maxima-Based Robust Digital Image Watermarking in Wavelet Domain

A new robust watermarking approach was proposed in 2D continuous wavelet domain （CWT）. The watermark is embedded into the large coefficients in the middle band of wavelet transform modulus maxima （WTMM） of the host image. After possible attacks, the watermark is then detected and extracted by correlation analysis. Compared with other wavelet domain watermarking approaches, the WTMM approach can endow the image with beth rotation and shift invariant properties. On the other hand, scale invariance is achieved with the geometric normalization during watermark detection. Case studies involve various attacks such as shifting, lossy compression, scaling, rotation and median filtering on the watermarked image, and the result shows that the approach is robust to these attacks.

基于DDS和框架的鲁棒盲水印算法

为进一步提高盲水印算法的鲁棒性,提出了一种基于DWT-DCT-SVD和框架的鲁棒盲水印算法。首先提取彩色宿主图像的Y、U、V分量灰度图,分别进行一级小波变换,对低频子带4×4分块并进行离散余弦变换和奇异值分解;然后通过调节每个子块的对角矩阵,将置乱的水印信息嵌入到宿主图像中;最后在图像四周添加框架用以定位和几何校正。实验结果表明,嵌入水印后图像SSIM和PSNR值分别达到0.96和39.06以上,攻击后提取的水印图成功跳转率在90%以上。该算法在不影响原图视觉效果的前提下,不仅提升了盲水印算法对于旋转、缩放、噪声等常规攻击的鲁棒性,还能抵抗任意角度旋转、大尺度缩放和截图的联合攻击。

基于随机分布圆形模板的抗屏摄盲水印算法

针对现有的抗屏摄水印算法不能兼顾鲁棒性、不可见性和水印容量等问题,提出一种基于随机分布圆形模板的抗屏摄鲁棒盲水印算法。首先将十六进制水印序列转换为固定的角度值,再根据随机线条分布算法生成一系列带有角度信息的圆形模板,将原宿主图像蓝色分量进行整数小波变换并取出低频子带(LL),对LL分块并用密钥选出若干水印嵌入块,最后将圆形模板嵌入被选出的块中。另外,为了提高水印容量,选择十六进制水印密码以代替二维码或条形码;为了提高水印不可见性,将水印模板嵌入整数小波变换的低频子带;为了平衡水印透明性和鲁棒性,利用粒子群寻优算法(PSO)自适应地得出最佳嵌入强度;为了解决拍摄后造成的图像畸变,在水印提取前对拍摄的图像进行几何校正。在数据集USC-SIPI上选取了测试图,进行了常规攻击与屏摄攻击实验,并与该领域主流的六种方法进行了比较。实验结果表明,在保证不可见性的前提下,该算法能抵抗各种常规类型攻击。在一定范围的拍摄距离和拍摄角度内,算法能有效提取出水印密码,并提高了水印容量,在抗屏摄方面具有良好的鲁棒性。

基于DCT的暗水印技术的研究与应用

各大公司或者企业越来越注重内部电子文件资料的安全性问题。针对电子文件形式的资料泄露问题,研究了一种基于DCT变换的数字水印技术,开发了一款辅助用户添加与提取暗水印的系统。该系统使用了Hough变换与Arnold置乱算法,提供了一种生成不严重干扰用户视觉,又能够抗提取、抗旋转、抗裁剪等一系列攻击的暗水印应用方案,该方案在保障公司内部电子文件不泄露的同时也兼顾了个人信息的安全性问题,满足了实际生产环节中公司或者企业保护电子文件的需求。

基于卷积和抖动调制的音频数字水印算法

为增强鲁棒性和透明性,提出一种基于卷积和抖动调制的音频水印算法。以浊音帧中能量最大的子帧为标记设计一种同步机制以自动定位水印所在位置。利用卷积运算构建一个具有稳定特性的特征量,通过量化和抖动调制将水印嵌入特征量。实验结果表明,该算法可实现水印的盲提取,能够抵御多种攻击类型,在86 bps的隐藏容量下信噪比可达26 dB以上,与同类算法相比具有更强的鲁棒性和更高的透明性,且隐藏容量越小,透明性和鲁棒性越好。

基于Amold加密与GBT-DWT的数字盲水印算法研究

为维护图像版权,实现对数字作品的保护,文章研究了一种基于Amold加密与GBT-DWT的数字盲水印方法。该算法运用置换原理对具有识别信息的水印图片做Amold置换预处理,通过图像转换技术与离散小波二次变换逐步分解载体图片,最后将置乱水印嵌入到变换后的载体图片中,实现数字作品的版权保护。实验结果表明,嵌入的水印信息低于人类视觉系统所能承受的最小力度发现范围,经文章算法嵌入后的图像具有较好的鲁棒性与隐藏性。

基于H.264低比特率视频流的半脆弱盲水印算法实现

本文提出一种基于H·264低比特率视频流的半脆弱盲水印方法,通过在H·264特有的的帧内预测和运动向量预测中分别嵌入鲁棒水印和脆弱水印,达到版权保护和内容完整性认证的双重目的.通过在嵌入强度和预测模式选择的拉格朗日乘子中引入H·264量化因子,增强了水印对重量化编码的抵抗力,取得更好的率失真平衡,减小水印嵌入对视频流比特率的影响.该算法可以实现水印的快速嵌入提取,满足视频实时处理的要求,实验结果证明了该算法的有效性.

一种基于直方图平移的高性能可逆水印算法

近年来,基于直方图平移的可逆水印算法因其生成水印图像质量高而被广泛研究.目前大部分相关水印算法是非盲的,需要额外传送峰值/零点对等边信息.已有的相关盲水印算法由于采用固定的峰值/零点对而性能较差.本文提出一种面向差值域多层嵌入的高性能直方图平移盲可逆水印算法,有关同步机制可保证在每层嵌入时都选择最优的峰值点和零点.对于单层嵌入,提出两步嵌入法(Two-step embedding,TSE)同步机制,将边信息嵌入到第一步生成的水印像素最低有效位以实现盲提取;对多层嵌入,则进一步提出链式嵌入同步机制,并在最后层采用两步嵌入法以保证水印盲提取.仿真结果表明,本文算法不但可以实现盲提取的要求,而且在算法性能方面较同类算法有显著的提高.

基于DWT-SVD压缩量化的数字图像盲水印算法

为了解决奇异值分解(singular value decomposition,SVD)算法提取水印时需要原始载体图像的缺陷,以及量化索引调制(quantization index modulation,QIM)均匀量化不适用于非均匀信号的问题,通过引入μ律压缩技术,提出一种新的基于DWT-SVD压缩量化的数字图像盲水印算法。该算法对载体图像进行分块,对每一分块实施离散小波变换(discrete wavelet transform,DWT),以及对变换后的近似部分系数进行SVD分解,使用μ律压缩函数压缩分解后的最大奇异值,用QIM的方法嵌入二值水印。算法只用到了最大奇异值,可以盲提取水印,消除因传输原始载体图像产生的不安全性,μ律压缩技术也减小了嵌入水印对原始载体图像的扰乱。仿真实验结果表明,该算法保持了较高的透明性,并对高斯噪声、中值滤波、联合图像专家小组(joint photographic experts group,JPEG)压缩、缩放等常见攻击具有更强的鲁棒性。

基于坐标映射的矢量地理数据全盲水印算法

通过对全盲水印特点的分析,提出了将文字按照字符编码方式直接转化为二进制位的水印信息生成方式。基于矢量地理数据的空间定位特性,研究了水印嵌入与提取中坐标映射机制和映射函数构造原则。利用提出的水印生成和坐标映射方法,建立了一种矢量地理空间数据全盲水印算法。实验结果表明,提出的算法能抵抗矢量数据中常见的数据压缩、增点、删点、编辑、裁剪以及平移等攻击,具有很好的鲁棒性。

基于SVD的图像数字水印算法研究

数字水印技术被视为数字产品版权保护的一种新方法。利用奇异值分解(singularvaluedecomposition,SVD)的良好性能以及双密钥调制的特点,提出了一个新的基于SVD的数字图像水印算法。首先将图像进行分块,然后通过各个图像子块的SVD变换获得奇异值,将各子块的奇异值组成向量后再采用双密钥调制方式嵌入水印。水印的提取不需要原始图像。实验结果表明该算法具有良好的鲁棒性和不可见性。

基于常函数的抗几何变换的矢量地理数据水印算法

基于常函数对于某些变换具有的不可变性,对抗几何变换的矢量地理数据水印算法进行研究。首先,对常函数的概念及其特征进行阐述,并推证一个抗旋转、平移和放缩攻击的常函数;然后,基于建立的常函数,通过把水印信息嵌入到几何不变量中,提出一种抗几何变换的矢量地理数据水印算法;最后,对提出的水印算法的鲁棒性进行试验分析。试验结果表明,提出的算法能抵抗压缩、增点、删点、裁剪、平移、旋转、放缩以及由以上攻击组成的复合攻击,具有较好的鲁棒性。