基于图像仿射不变特征点的零水印算法

本文提出了一种基于图像提升小波变换和仿射不变特征点的零水印算法。首先对图像进行3级提升小波变换,然后在低频部分,利用Harris-Affine算子提取出图像的仿射不变特征点和仿射协变特征区域,利用这些特征区域来构造零水印信息。实验证明,该算法不仅对常规信号处理攻击和简单的几何攻击具有很强的鲁棒性,对于剪切、纵横比改变、行列去除,局部扭曲等较复杂的几何攻击也具有较强的抵抗能力。     

基于SIFT特征点的抗几何攻击水印算法

提出一种利用尺度不变特征变换（SIFT）特征点实现对抗几何攻击的水印方案。基于特征点构造Delavnay三角网,依靠SIFT特征点的抗几何变换性实现图像三角形区域的抗几何攻击。在提取部分图像内容的基础上,按照多对一的原则,使用改进的加性方法嵌入水印信息。实验表明,该算法能够较好地抵抗几何攻击和一般的信号处理攻击。     

基于仿射协变区域的抗几何攻击图像水印算法

提出了一种基于仿射协变区域的抗几何攻击数字图像水印算法. 首先利用仿射不变点检测算子提取图像特征点, 并通过基于最小生成树聚类算法的选择策略获得一组稳定且彼此独立的仿射协变特征区域. 然后基于特征区域, 利用图像归一化和主梯度方向对齐得到具有缩放和旋转不变性的圆形区域. 最后将水印在空域直接嵌入到圆形区域中. 大量的实验结果表明本文方法是有效的, 对常规图像处理、几何攻击以及组合攻击均具有较好的鲁棒性.     

基于图像特征的鲁棒水印算法研究

为提高图像盲水印抗几何攻击的能力,提出了一套基于图像特征的水印嵌入方案。首先利用具有几何不变性的Harris算子检测出图像特征点,然后在每块由特征点生成的Delaunay三角网上采用一种基于小波系数显著性差异调制量化方式进行水印的嵌入和检测。该嵌入水印算法的思想是通过比较每组小波系数的显著性差异值和所有组的平均显著性差异值的方法来量化每一组块的最大小波系数。实验结果表明,该方法不仅对压缩、滤波等常见攻击具有稳健性,同时对旋转、缩放、变形等几何攻击也具有较好的稳健性。     

Rich-information watermarking scheme for 3D models of oblique photography

The security issues of copyright information arise rapidly along with the widely application of 3D models of oblique photography. In this paper, a reversible watermarking scheme is proposed for 3D models of oblique photography, aiming at robustness enhancing and rich-information watermark embedding. To realize the goal of rich-information watermark embedding, an encoding method of original copyright information is designed based on QR code, which can shorten significantly the length of the final watermark data without affecting the copyright expression. A 3D point grouping method is designed and both watermark embedding and extraction are completed group by group, so that the robustness of this scheme under certain attacks can be enhanced, e.g., model segmentation, data compression, point randomly deleting, etc. Moreover, watermarks are embedded into spherical coordinates of 3D points so that the ability of the proposed scheme in resisting geometric transformation can be improved. Results of simulation experiments have demonstrated that the proposed scheme can support rich-information watermark embedding, and it has a satisfying robustness under common geometric and non-geometric attacks.

     

基于SIFT的椭圆区域鲁棒数字水印方案

如何有效抵抗去同步攻击是数字图像水印研究领域的热点问题之一。利用图像仿射协变特征,提出一种可有效抵抗去同步攻击的鲁棒水印算法。对目前流行的利用多尺度Harris和SIFT描述算子来匹配图像的方法,后者有较好的匹配效果,对恢复同步水印更加稳定,并且能较好抵抗去同步攻击。该算法利用性能稳定的SIFT算子提取图像特征点,并通过基于最小生成树聚类算法的选择策略获得一组稳定且彼此独立的椭圆仿射协变特征区域,基于特征区域,利用椭圆归一化得到具有缩放和旋转不变性的圆形区域。将圆形区域进行非下采样轮廓变换（NSCT）,其中非下采样轮廓变换不仅克服了小波（Wavelet）变换的非奇异性最优基缺点,而且提供了优于轮廓（Contourlet）变换的平移不变性。将水印嵌入变换后的低频子带中。该算法实现盲提取,仿真实验结果表明,提出的算法是有效的且对常规图像处理、几何攻击以及组合攻击均具有较好的鲁棒性。     

具有抗几何攻击能力的盲数字图像水印算法

几何攻击作用在隐秘图像上,这导致检测水印失败,除非在检测水印之前能使受到几何攻击的隐秘图像恢复到与原始图像具有同样大小和位置,也就是恢复已失去的同步信息.为了做到这一点,解码器需要知道隐秘图像受到了怎样的几何攻击.文章提出了一种新的估计几何变换的方法,它利用图像的形殊点来实现这个目的.分别计算出原始图像和测试图像的形殊点集,通过模拟退火算法对两个形殊点集进行匹配,从而估计出几何变换的参数.所提出的方法在检测水印时不需要原始图像做参考,这是因为原始图像的形殊点信息已包含在密钥中.实验结果表明该方法对几何攻击和其它常用图像处理攻击具有鲁棒性.     

Robust image watermarking technique using triangular regions and Zernike moments for quantization based embedding

Watermarking is a tool to embed information in the image to provide authentication, copyrights protection, copy control, etc. Some watermarking techniques are robust to intentional /unintentional attacks on the watermarked image. In this study, we propose a robust watermarking approach that can resist geometrical attacks. The proposed technique exploits both the robust image feature points and local Zernike moments for embedding the information. Delaunay tessellation is employed to divide image into distinct triangular segments based on robust features. These features are identified using Harris detector. Zernike moments are calculated for each selected triangular segment, and then the watermark is embedded in the magnitude of Zernike moments using dither modulation. It can be observed from the experimental results that by using proposed approach, the watermark can be detected even in the presence of geometrical distortion, i.e. rotation, cropping, and scaling, and JPEG compression attack.

     

Robust localized image watermarking based on invariant regions

The robustness of the localized watermarking methods mainly depends on the robustness of the feature locating the watermark. Based on the mean luminance of the disk, a rotation and scale invariant feature extraction algorithm is proposed. A theoretical verification of the rotation and scale invariance of extracted feature points in the continuous image is further performed. The extracted feature points are used to construct rotation and scale invariant circular regions, where the watermark is embedded after affine normalization. Experimental results show that the constructed regions fit the watermarking applications much better than those in previous feature-based watermarking schemes from the aspect of robustness against common attacks including filtering, JPEG compression, cropping, rotation and scaling, and the proposed localized image watermarking scheme has better robustness than previous feature-based watermarking schemes against common signal process and geometrical attacks while maintaining imperceptibility.     

一种基于特征点的三维网格盲水印算法

针对三维网格水印空域算法无法兼顾嵌入量与透明性,且水印盲检测过程繁琐的问题,提出一种基于特征点的盲水印算法。选取三维模型最远两点为全局特征点建立全局坐标系,根据仿射不变性原理将原始载体仿射到一个固定的球型空间,以增强对多种几何攻击的抵抗能力。根据载体点密集程度将其等分成若干个局部空间,以局部空间中离质心最远点为局部特征顶点建立局部几何坐标系,从而增强算法的透明性。利用顶点在坐标系的投影角度来存储水印索引,实现盲水印。实验结果表明,该算法能有效抵抗旋转、缩放、噪声等攻击,具有强鲁棒性和较好的不可感知性,同时兼具盲水印检测优势。     

基于特征点的抗几何攻击零水印

提出了最小偏离度的概念,根据IPR库中的特征点信息和待测图像自身的特征点信息分别构造Delaunay三角网,并以此计算两张三角网中三角形的最小偏离度值。若该值低于预先设定的阈值,则可以确认待测图像的版权归属。实验表明,该算法对各种几何攻击甚至是组合几何攻击均具有较强的鲁棒性。     

基于第二代Bandelet变换的抗几何攻击图像水印

抗几何攻击的鲁棒图像水印设计是目前水印技术研究的难点和热点之一. 文中分析了图像的Bandelet变换特性, 提出了一种以图像特征点矢量集为特征向量的回归支持向量机(Support vector regression, SVR)和第二代 Bandelet变换的抗几何攻击图像水印算法,采取的主要方法包括: 1)在Bandelet变换提取的刻画图像局部特征的几何流系数上, 采用奇偶量化嵌入水印; 2)利用Harris-Laplace算子从归一化的含水印图像中提取具有几何形变鲁棒性的图像特征点,构造特征点矢量集 作为特征向量,应用回归支持向量机对几何变换参数进行训练学习; 3)水印检测时, 先利用SVR训练模型得到待检测图像所受几何攻击的参数并作几何校正,然后通过奇偶检测器盲提取水印.仿真实验表明,所提出的水印算 法不仅具有良好的透明性,而且对常规图像处理、一般性几何攻击和组合攻击均具有良好的鲁棒性.     

DWT域抗几何攻击水印算法研究

鲁棒水印面临的最大问题就是几何攻击.小波变换域不具有几何不变性,一个微弱不可察觉的几何攻击就能使小波域水印算法失效.提出了一种新的以特征点作为模板、有效抵抗几何攻击的图像水印算法.它包括:(1)根据小波变换空间各子带呈树形结构的特性,选择每棵小波树中纹理最强的方向子树的树根作为水印嵌入点;(2)根据嵌入点在低频对应位置的能量和该方向子树高频叶子结点的纹理信息确定一种自适应嵌入水印策略;(3)利用Harris-Laplace算子从含有水印的图像中提取出具有几何形变鲁棒性的特征点,将其作为模板;(4)水印检测时,先利用特征点模板通过线性变换恢复几何形变的图像.然后通过统计图像的相关性来验证水印的存在,无需原始图像.仿真实验表明,所提出的水印算法不仅具有很好的透明性,而且对常见的图像处理和几何变换具有很好的鲁棒性.     

局部特征点的3维模型变换域水印算法

目的 为进一步提高变换域水印的鲁棒性,解决多分辨率小波变换对三角形网格形状的限制,提出一种基于局部特征点的变换域水印算法。方法 首先根据经度角和纬度角对模型分区,并求出局部特征点;其次由局部特征点的模长与非局部特征点模长的均值构成的比值组成2维矩阵,对2维矩阵进行离散小波变换,并修改小波系数中的各分辨率层的中高频系数来嵌入水印;最后通过逆向离散小波变换得到空域信号,修改3维模型局部特征点的模长以得到含水印的3维网格模型。结果 本文方法能够不受平移、旋转、均匀缩放和顶点重排序的影响,并能有效抵抗噪声、简化、剪切、细分、量化和联合攻击等一系列的常见攻击。结论 实验结果表明,该算法具有很好的鲁棒性和不可见性,对模型的网格形状无严格要求。     

Contourlet-based image and video watermarking robust to geometric attacks and compressions

In this paper, we first propose a new blind image watermarking scheme robust to geometric attacks and compressions. The scheme is based on contourlet transform (CT) and principal component analysis (PCA). The scheme uses the principal components of the largest contourlet coefficients of the last directional subband of the cover image to embed the watermark. Meanwhile, with the noise visibility function (NVF), the watermarking strength is adjusted adaptively to preserve the perceptual quality of the image. The watermark can be detected with high accuracy after various possible distortions. The normalized correlation (NC) between the original watermark and the watermark extracted from the distorted watermarked image is used as the robustness evaluation criterion. The simulation results demonstrate that the proposed scheme has good performance in terms of both quality and robustness against a variety of image-processing attacks, such as rotation, scaling and image compressions. Then we extend the scheme to blind video watermarking. The performance of the video watermarking scheme is evaluated against video attacks like rotation, frame averaging, noise additions and video compressions. The introduction of the CT produces robustness against image and video compressions, and the PCA yields resistance to geometric attacks.

     

基于特征点模板的Contourlet域抗几何攻击水印算法研究

抵抗几何攻击的数字水印是目前水印算法研究的热点与难点之一,一个微弱不可察觉的几何攻击就可能使绝大多数水印算法失效.文中对图像的Contourlet变换特性进行了分析和讨论,在此基础上提出了一种以特征点为模板的基于Contourlet的抗几何攻击水印算法,包括:(1)在Contourlet变换域最高阶方向子带的相同带内坐标中搜索纹理最丰富的位置作为嵌入位置,这样水印在被嵌入到图像重要区域的同时也保证其被嵌到纹理丰富的区域中;(2)根据嵌入点在低频对应位置的能量和高尺度同方向子带对应位置的纹理信息,确定了一种自适应的水印嵌入强度的策略,使得水印在嵌入位置和嵌入强度的选择上都保证了嵌入的水印具有很好的透明性和极大的鲁棒性;(3)改进了Harris-Laplace算子,利用该算子从含有水印的图像中提取出具有几何形变鲁棒性的特征点作为模板;(4)水印检测时,先利用特征点模板对应点之间的变换关系实现水印重同步,然后通过统计图像的相关性来验证水印的存在.仿真实验表明,所提出的水印算法不仅具有很好的透明性,而且对常见的图像处理和几何攻击具有很好的鲁棒性.     

Color Image Zero-Watermarking Using Accurate Quaternion Generalized Orthogonal Fourier–Mellin Moments

In recent years, digital image zero-watermarking algorithm has made great progress. But there are still some problems. First, most algorithms only focus on robustness and ignore discrimination. Second, most methods have good robustness against conventional signal attacks but poor robustness against geometric attacks. Thirdly, most of the existing zero-watermarking algorithms only focus on gray-scale images, and there are relatively few researches on color images. In order to cope with these issues, this paper introduces a novel color image zero-watermarking scheme using accurate quaternion generalized orthogonal Fourier–Mellin moments (AQGOFMMs). In the first stage of the proposed method, accurate computation of generalized orthogonal Fourier–Mellin moments (GOFMMs) based on the polar pixel tiling scheme. In the next stage, the high-precision GOFMMs are extended to the accurate quaternion GOFMMs (AQGOFMMs). Finally, the immutable set of features is extracted to construct a zero-watermark. It is worth mentioning that, different from the traditional way of constructing watermark with the amplitude of moments, this algorithm uses the full 4-D features of AQGOFMMs to construct watermark. The experiment proved that proposed zero-watermarking scheme gives a good balance between discriminability and robustness. Furthermore, the proposed algorithm achieves better performance than the other existing zero-watermarking.

     

A robust video zero-watermarking based on deep convolutional neural network and self-organizing map in polar complex exponential transform domain

In this paper, a robust video zero-watermarking scheme for copyright protection using a combination of convolutional neural network (CNN) and self-organizing map (SOM) in polar complex exponential transform (PCET) space is presented. The scheme is developed not only to remedy the existing problems of lacking in some performance assessments but also to enhance the robustness. It starts with extracting the content feature of each frame by CNN and then some significant frames are selected using SOM clustering and maximum entropy. Secondly, the PCET is applied to all selected frames to abstract invariant moments, and further, is scrambled by a chaotic logistic map and is reduced in dimensions by singular value decomposition (SVD). Next, a binary sequence is generated by comparing adjacent values of the obtained compact PCET moments in the previous step, and further is permuted to produce a binary matrix. Finally, a bitwise exclusive-OR operation is imposed on the binary matrix and the encrypted watermark by the chaotic map to generate a zero-watermark signal. Experimental results demonstrate that the proposed scheme has adequate equalization and distinguishability of zero-watermarks as well as strong robustness against common signal processing, geometric, compression, and inter-frame attacks. Also, compared with existing video zero-watermarking and traditional video watermarking methods, the proposed scheme exhibits superior robustness.

     

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     

基于仿射变换和线性回归的3D人脸姿态估计方法

提出了一种由仿射变换关系到线性回归的3D人脸空间姿态估计方法。即跟踪到人脸特征点后,根据仿射变换关系得到人脸姿态的粗估计值,以这个粗估计值作为人脸姿态的初始值,再通过线性回归迭代求得人脸姿态的精确值。实验结果表明,该方法在较大的姿态变化范围内,具有良好的估计精确度和鲁棒性。