基于标记矢量量化的多重音频数字水印算法

提出一种基于矢量量化和DCT变换的多重音频数字水印算法.该算法将鲁棒水印和脆弱水印分别嵌入到载体音频信号中,同时完成版权保护和音频内容完整性认证的功能.鲁棒水印在进行码字标记的矢量量化过程中嵌入到DCT中频系数中,脆弱水印采用索引约束的方法嵌入到DCT高频系数中.鲁棒水印和脆弱水印都能实现盲提取.实验结果表明,本算法引...     

基于高斯混合模型的DCT域水印检测方法

基于音频DCT系数的统计特征,提出了一种水印检测方法.采用扩频的方法在DCT域嵌入水印, 分别对嵌入水印和未包含水印的音频信号在DCT域进行统计学习,得到对应的高斯混合模型(Gaussian mixture model, GMM). 接收端采用最大似然检测,判断是否嵌入水印并提取相应的水印信息. 仿真结果表明本文的水印检测算法对常见的信号攻击具有鲁棒性,与传统的相关检测法相比,检测可靠性更高.     

基于Logistic混沌序列的鲁棒数字水印方案

数字图像水印研究中,针对图像水印鲁棒性和隐蔽性的矛盾、数字水印本身的安全问题,结合混沌系统原理,提出一种Logistic双混沌系统的鲁棒数字图像水印方案。不同于许多传统水印算法,水印嵌入时不将原始水印嵌入,而是根据密钥生成两个混沌序列作为水印序列,然后按照规则,依次根据原始水印的每个比特值选取某个混沌水印序列的一比特信息嵌入载体图像对应的DCT中频系数中。在提取水印时,根据DCT的中频系数和混沌水印序列的相关性来确定每一位原始水印比特值,最终合并得到完整的水印图像。由于水印序列的产生过程经过了混沌加密处理,所以算法效率较好,且保证了水印本身的安全性。仿真结果表明,水印嵌入和提取算法都具有良好的效果,并且具有较强的抗攻击性能。     

数字音频信号的脆弱水印嵌入算法

该文给出了量化系数嵌入水印比特信息的基本原理，在此基础上提出了数字音频信号的脆弱水印嵌入算法。该算法充分利用离散子波变换的多分辨率特性，通过等概率随机量化音频信号不同子带的子波系数嵌入视觉可辨别的水印－二值图像。仿真实验结果表明，该文提出的数字音频信号脆弱水印嵌入算法对有损压缩、滤波和重新采样等操作具有很强的敏感性。     

基于心理声学模型的自适应量化同步数字音频水印算法

针对传统的固定步长量化方法对不同种类音频载体适应性较差的缺点,基于心理声学模型,在获取的语音信号的DCT域的低频系数中量化地嵌入水印信息,量化步长由掩蔽阈值自适应控制,以满足水印的盲检测。同时引入了同步信号的思想,利用同步信号和水印序列号定位水印的嵌入位置。实验结果表明,该算法不仅具有较好的不可感知性,而且对诸如低通滤波、有损压缩、重采样和剪裁等攻击均具有较好的顽健性。     

基于零系数索引的可逆图像水印

提出一种新的用于压缩图像的具有高容量和低失真特点的可逆水印算法,在提取水印后可还原原始的压缩图像.该算法使用量化DCT块的中高频零系数索引值进行水印嵌入和提取.只要候选系数范围内最大的零系数索引不小于水印的整数值,就可以通过修改一位零系数来嵌入多个比特的水印.由于降低了被修改系数与嵌入的水印位数比例,该算法能在引入较少失真的情况下嵌入大量的水印.实验结果表明相对于其他基于DCT压缩域的可逆图像水印算法,该算法在水印容量和嵌水印图像质量方面具有优势,并且具有良好的适应性和可调整性.     

一种新的基于量化的音频数字水印算法

文章提出了一种新的基于量化的音频数字水印算法,该算法首先将视觉上可辨的二值水印图像降为一维序列,再将水印序列伪随机排序,最后对数字音频信号进行分段DCT(离散余弦变换),水印信号通过量化处理过程嵌入到DCT变换后的系数中,提取水印信号无需使用原始数字音频信号。仿真实验表明水印是不可察觉的,经过mp3有损压缩、低通滤波、重采样等操作后仍具有很强的稳健性。     

一种DCT域自适应音频水印算法*

数字水印技术中水印强度是影响水印鲁棒性的重要参数。当前大多数的水印算法均利用实验来确定水印强度,但实验具有很大的随机性,得到合适的水印强度需要大量的实验。提出了一种DCT(discrete cosine transform)域基于信噪比(signal noise ratio,SNR)水印强度自适应的音频水印。首先将载体音频分帧,然后根据对嵌入水印后音频品质要求的SNR计算各帧的水印强度,选择可嵌入强度大的帧,将一个水印比特信息嵌入一帧的DCT的直流系数,水印强度将自适应于各帧。实验结果证明了该算法具有较好     

一种基于图像矢量量化压缩的数据隐藏方法

根据码书中所有码字之间的最近邻关系,采用最近邻方法标记码书使得所有码字与其最近邻码字的标记值不同。在此基础上,运用I/矩阵运算可在矢量量化压缩图像的每n=2m个码字索引中有选择地最多修改m2」个从而嵌入m比特数据。计算机上的仿真实验结果表明：该方法不仅具有较大的嵌入容量,而且具有较好的隐蔽性,可用于矢量量化压缩图像中的信息隐藏。     

基于DWT-SVD的奇异向量量化水印算法

提出了一种新颖的可用于版权保护的小波奇异值分解的量化水印算法。与传统的水印比特信息直接嵌入小波系数不同,水印信息被量化嵌入原始图像小波低频子带分块奇异值分解得到的奇异向量中。水印提取无需原始图像,可在密钥和量化阂值控制下实现盲提取。实验表明,含水印图像质量好且能较好地抵抗常规的图像处理,对JPEG压缩具有优异的鲁棒性。     

A semi-fragile watermarking scheme for authenticating audio signal based on dual-tree complex wavelet transform and discrete cosine transform

In this paper, a novel semi-fragile watermarking scheme for authenticating an audio signal based on dual-tree complex wavelet transform (DT-CWT) and discrete cosine transform (DCT) is proposed. Specifically, the watermark data are efficiently inserted into the coefficients of the low-frequency sub-band of DT-CWT taking advantages of both DCT and quantization index modulation (QIM). First, the original digital audio signal is segmented and then performed with DT-CWT. Second, based on the energy compression property, the low-frequency sub-band coefficients of the DT-CWT domain are performed with DCT, and the DC component is utilized to embed one distorted watermark bit by the QIM technique. Finally, inverse DCT and DT-CWT are orderly implemented on the watermarked coefficients of each audio segment to get a watermarked audio signal. Simulation results show that the hybrid embedding domain constructed by DT-CWT and DCT is effective, and the proposed watermarking scheme is not only inaudible, but also robust against content persistent non-malicious audio signal processing operations, such as MP3 compression, noise addition, re-sampling, re-quantization, etc. Furthermore, the proposed scheme can effectively authenticate the veracity and integrity of audio content and greatly expands the applicability of the audio watermarking scheme.     

Adjustable audio watermarking algorithm based on DWPT and psychoacoustic modeling

This paper presents a novel adjustable audio watermarking method with high auditory quality by exploiting the discrete wavelet packet transform (DWPT), psychoacoustic modeling and distortion compensated-dither modulation (DC-DM) quantization. While the DWPT is used to divide the audio frames into several frequency sub-bands, the psychoacoustic model is intergraded to determine the appropriate sub-bands for watermarking and to control the number of embedded bits in each one. Then, the DC-DM technique is used to embed the watermark bits into the appropriate DWPT coefficients. The synchronization code technique is adopted in the proposed method to withstand desynchronization attacks. In order to achieve an adjustable watermarking scheme, two regulator parameters are provided to manage the capacity-robustness trade-off. The performance of the watermarking scheme is evaluated by examining different host audio signals under various watermarking attacks. The results show excellent imperceptibility of watermarked signals with an average ODG of ? 0.3. In addition, the proposed scheme provides strong robustness against the attacks with low capacity. However, high capacity (about 2500 bps) can be achieved while maintaining a reasonable robustness. A comparison with some state-of-the-art audio watermarking schemes reveals that the proposed method provides competitive results.     

用于MP3作品版权保护的数字水印新算法

以MP3压缩攻击特点与听觉掩蔽特性为基础，提出了一种可用于MP3音乐作品版权保护的混合域数字音频盲水印算法．首先，结合MP3压缩等攻击特点，对原始数字音频进行自适应的分段处理；再对每个音频数据段进行DWT和DCT；最后，结合听觉掩蔽特性与MP3压缩攻击特点，自适应确定量化步长，并将水印信号嵌入到音频数据段的低频分量中．仿真实验表明，所提出的算法不仅具有较好的透明性，而且对诸如重新采样、重新量化、叠加噪声、低通滤波等攻击均具有较好的鲁棒性．特别地，该算法对于最为普通的MP3压缩攻击具有极强的抵抗能力（尤其是高压缩比下）．     

Variable-dimensional vector modulation for perceptual-based DWT blind audio watermarking with adjustable payload capacity

A variable-dimensional vector modulation (VDVM) scheme is introduced to maximize the efficiency of the norm-space DWT-based blind audio watermarking technique. This flexible scheme allows the watermarking algorithm to reach a balance between robustness and capacity, while the imperceptivity is always ensured. The imperfection of applying quantization index modulation in the open-loop case has been rectified. The effectiveness of the proposed scheme is proven using the perceptual evaluation of audio quality (PEAQ) and bit error rates of recovered watermarks under various signal processing attacks. Experimental results show that the proposed VDVM scheme is comparable to other recently developed methods in robustness and imperceptivity even at a capacity as high as 301.46 bps. Such a capacity can be further doubled by halving the dimension of the involved DWT vector, while the robustness is still maintained at a satisfactory level.     

An audio watermarking scheme using singular value decomposition and dither-modulation quantization

Quantization index modulation is one of the best methods for performing blind watermarking, due to its simplicity and good rate-distortion-robustness trade-offs. In this paper, a new audio watermarking algorithm based on singular value decomposition and dither-modulation quantization is presented. The watermark is embedded using dither-modulation quantization of the singular values of the blocks of the host audio signal. The watermark can be blindly extracted without the knowledge of the original audio signal. Subjective and objective tests confirm high imperceptibility achieved by the proposed scheme. Moreover, the scheme is quite robust against attacks including additive white Gaussian noise, MP3 compression, resampling, low-pass filtering, requantization, cropping, echo addition and denoising. The watermark data payload of the algorithm is 196 bps. Performance analysis of the proposed scheme shows low error probability rates.     

基于经验模式分解的音频水印算法

针对传统基于经验模式分解(EMD)的音频水印算法鲁棒性不强的问题,提出一种基于固有模态函数(IMF)极值的盲音频水印算法.首先对音频信号进行分帧,每个音频帧经过EMD后得到IMF; 接着利用均匀量化的方法将水印信息和同步码嵌入到最后一个IMF的极值中.所提算法的数据嵌入率是46.9~50.3 b/s,且携水印音频保持了原始音频的感知质量.对携水印音频进行加噪、MP3压缩、重新采样、滤波、剪切和重采样攻击后,提取出的水印信息变化不大,算法鲁棒性较好.与时间域和小波域算法相比,提出的算法在保证高数据嵌入率的同时,可以抵抗32 kb/s的MP3压缩攻击.     

A New Adaptive Digital Audio Watermarking Based on Support Vector Regression

On the basis of support vector regression (SVR), a new adaptive blind digital audio watermarking algorithm is proposed. This algorithm embeds the template information and watermark signal into the original audio by adaptive quantization according to the local audio correlation and human auditory masking. The procedure of watermark extraction is as follows. First, the corresponding features of template and watermark are extracted from the watermarked audio. Then, the corresponding feature of template is selected as training sample to train SVR and an SVR model is returned. Finally, the actual outputs are predicted according to the corresponding feature of watermark, and the digital watermark is recovered from the watermarked audio by using the well-trained SVR. Experimental results show that our audio watermarking scheme is not only inaudible, but also robust against various common signal processing (such as noise adding, resampling, requantization, and MP3 compression), and also has high practicability. In addition, the algorithm can extract the watermark without the help of the original digital audio signal, and the performance of it is better than other SVM audio watermarking schemes.     

DCT域音频水印算法的嵌入位置选择策略

本文在DCT最大误差模型研究的基础上提出了一种新的音频水印嵌入位置选择策略。建立了水印嵌入位置和嵌入水印后的音频文件的听觉感知性之阿的关系,根据音频水印的不可听性的要求选择最优的嵌入位置,然后选择不同的水印强度调整鲁棒性,从而最大程度地保持音频水印的不可听性和鲁棒性,为解决音频水印嵌入过程中不可听性和鲁棒性之间的矛盾提供了一种策略。     

A blind audio watermarking algorithm by logarithmic quantization index modulation

In this paper we present a blind audio watermarking algorithm based on the vector norm and the logarithmic quantization index modulation (LQIM) in the wavelet domain, integrating the robustness of the vector norm with the imperceptibility of the logarithmic quantization index modulation based on μ-Law (or mu-Law) companding. Firstly μ-Law companding is adopted to transform the vector norm of the segmented wavelet approximation components of the original audio signal. And then a binary image scrambled by the chaotic sequence as watermark is embedded in the transformed domain with a uniform quantization scheme. Experimental results demonstrate that even if the capacity of the proposed algorithm is high, up to 102.4 bps, this algorithm can still maintain a high quality of the audio signal, and achieve a better performance, such as imperceptibility, robustness and complexity in comparison with the uniform quantization based algorithms against common attacks. What’s more, it can resist amplitude scaling attack effectively.