基于局部随机性的YASS隐写分析方法

提出了一种针对YASS(Yet Another Steganographic Scheme)隐写算法的专用检测方法.根据B块的大小,分析8×8数据嵌入子块的位置,并提取子块DCT低频系数局部随机序列的4个统计量作为特征,结合Fisher线性分类器进行分类.针对YASS及其改进算法的实验表明:此方法能有效检测YASS隐写得到的含密图像,相比现有方法,其检测率有明显提高.     

基于隐写编码和Markov模型的自适应图像隐写算法

如何构造大容量、低失真和高统计安全的隐写算法一直是隐写研究的难点和热点.提出一种兼顾感知失真和二阶统计安全的自适应图像隐写算法设计思路.算法将载体各部分的平滑度引入隐写编码的生成过程,自适应地利用一簇隐写编码在载体各部分的合理运用降低载密图像失真度;在隐秘信息嵌入方式上利用基于Markov链模型的动态补偿方法提高载密图像统计安全性;算法对载体最低有效位和次最低有效位进行嵌入以保证嵌入容量.实验表明算法在相同嵌入量下相较双层随机LSB匹配算法以及仅使用一种隐写编码的算法,失真度更低且载体统计分布的改变更小,而在失真度和统计分布改变相近时嵌入容量更大.     

基于相邻灰度值对互补嵌入的LSB匹配隐写改进算法

LSB匹配隐写具有嵌入量大.视觉隐蔽性高的优势,但采用LSB 匹配隐写算法对于载密图像的灰度直方图有明显的平滑作用,因此攻击者可以基于直方图分析图像是否载密.通过研究LSB匹配算法对直方图产生影响的机理,提出一种基于相邻灰度值互补嵌入的LSB 匹配改进算法.该算法利用匹配像素灰度值加减1对直方图的影响具有互补性的特点,以相邻灰度值匹配像素对为对象进行成对嵌入,有效地保持了直方图特性,极大地提高了算法的抗统计分析性能.     

一种空域BPCS信息隐藏的改进算法

针对位平面复杂度分割密写易受复杂度直方图攻击的问题,提出了一种密写改进算法,即用类噪声块一半的像素嵌入秘密信息,另一半用改变像素值的方法来调整区域的整体复杂度,以达到抗攻击的目的。实验结果表明,改进的算法能很好的抵抗复杂度直方图攻击。     

Canny-Gauss通用域图像隐写算法

自适应图像隐写算法是一种以图像为载体,通过手工设计嵌入失真代价,指导隐写码在图像载体中嵌入秘密消息的信息隐藏算法.长期以来,这类算法将秘密消息尽可能隐藏在图像纹理更深更复杂的位置以对抗基于富特征的隐写分析检测.然而,伴随着深度学习在隐写分析领域的快速发展,人工设计的自适应算法受到严重挑战.此外,基于加性失真的隐写编码在嵌入消息时,复杂纹理向边界聚集所产生的统计异常问题也亟待解决.因此,本文总结了各类人工失真代价的优势和不足,归纳出当前自适应算法在空域的设计范式,并结合UNIWARD在各嵌入域的转换规则,提出基于嵌入失真代价ρ的通用域隐写转换公式.然后,从隐写嵌入失真代价与图像纹理稀疏关系的角度出发,以Canny算子划分纹理、Gauss模糊缩放轮廓、AutoML搜索阈值的方式,提出了一种通用域隐写算法Canny Gauss.实验结果表明,本文所提通用域隐写转换公式能够有效应用于现有主流算法.同时,在UNIWARD所有可行嵌入域中,本文所提算法表达出更高嵌入失真代价稳定性和隐写隐蔽性,在第三方权重加持下的深度隐写分析表现与UNIWARD相比至少提升2.6%、最高提升14.6%.这为自适应隐...     

基于冗余零系数消除的F5改进算法

F5隐写分析算法利用收缩现象,比较剪裁后图像直方图和估计载体图像直方图的差异进行检测。针对该算法的特点,提出改进的F5算法,通过部分消除冗余零系数,提高嵌入率、防止收缩现象的发生。在标准图像集上的实验结果表明,改进算法改善了含密直方图的收缩现象,各检测参数与原始载体图像差别较小,可以有效抵抗F5隐写分析算法。     

基于运动矢量直方图不变的数字视频隐写算法

针对现有基于运动矢量(MV)的数字视频隐写算法对载体直方图统计特性带来较大改变这一问题,提出一种基于运动矢量直方图不变的数字视频隐写算法。利用保持直方图数据映射方法,把秘密信息隐藏在视频运动矢量中;同时,利用数据匹配编码对嵌入之前的秘密信息进行编码处理,得到了与视频运动矢量统计特征基本匹配的数据流,使算法接近于信息论下的完美安全。实验结果表明：算法有效控制了运动矢量直方图的改变,同时码率的增加也被有效地控制在1%以内,隐写分析检出率平均下降了30%~50%。     

一种基于像素预测的图像可逆信息隐藏策略

基于像素点预测的可逆信息隐藏(Reversible Data Hiding,RDH)是当今一种低失真、高容量的信息隐藏策略,特别是对于差值扩展和直方图平移算法而言,准确预测可以同时提高数据容量和减小图像失真。文中提出了一种基于像素预测差的直方图平移方案。首先,采用Warped Distance算法来对像素值进行预测,并在此基础上利用图像的局部梯度来实现对像素值的更精确预测。综合上述两种策略,改进了现有的直方图平移算法,同时也给出了防止像素值在平移后溢出的建议。实验结果表明,相对于近年来的其他方案,该方案能够在保证图像质量的情况下有效提升数据嵌入量,并且通过调整数据嵌入层级,可以在具体使用该算法时针对图像质量和数据嵌入量进行权衡。这也进一步说明了利用图像的局部特性,特别是梯度与图像局部几何相似性,可以有效提升像素预测的精度,从而改善可逆信息隐藏的容量-失真性能。     

用于实现医学图像ROI对比度增强的可逆数据隐藏算法

基于可逆数据隐藏技术具有可以准确地提取嵌入的秘密消息并无损还原原始载体的特点,可逆数据隐藏技术被广泛使用在医学和军事等领域。提出一种新的基于直方图平移的可逆数据隐藏算法,针对医学图像的像素分布特点,利用最大类间方差法分离背景和前景区域,确定边缘像素。对已有算法的预处理过程进行改进,使其能够自适应地选择最优预处理区间,降低视觉失真,并在信息嵌入的同时使医学图像感兴趣区域的对比度增强效果得到逐步改善,最终原始医学图像可以无失真地恢复。与现在已知的方法相比,新算法可以使医学图像ROI的对比度增强效果得到更明显的改善,大幅度提高了图像的视觉质量。     

基于LSBMR图像边缘自适应隐写的检测

在图像边缘自适应LSB匹配改进隐写算法中,秘密信息嵌入位置的选择仅由某个方向上像素对的差值决定,未考虑该像素与其邻域内其他像素的差值变化的特点.针对该问题,对隐写前后图像的八方向差分直方图进行分析,提出一种基于LSB匹配改进算法(LSBMR)边缘自适应隐写检测的算法.该算法计算图像的八方向绝对差分直方图,提取直方图中隐写前后变化较为明显的频数用以构建特征向量,并使用支持向量机完成检测.对较低嵌入率下(≤0.5 bpp)的EALSBMR隐写结果进行检测,结果表明该算法的平均检测率均高于现有典型的隐写分析算法.     

A channel selection rule for YASS

In this paper, a channel selection rule for YASS （Yet-Another-Secure-Steganography） is proposed. Secret message embedding imposes distortion to the cover image. The larger the distortion, the less secure the steganographic algorithm. Our channel selection rule engages in minimizing this distortion brought in by YASS. In our rule, the distortion caused by unit change on each quantized DCT （Discrete Cosine Transformation） component is computed, and the components with smaller unit change distortion are selected with higher priority. This channel selection rule reduces distortion to the medium spatial domain image and the final JPEO image. Experimental results show that our improved YASS scheme outperforms original YASS scheme on the aspects of both perception and statistics. This new channel selection rule can also be combined with other enhancements in YASS framework to further boost the performance.     

Steganalysis of YASS

A promising steganographic method—Yet Another Steganography Scheme (YASS)—was designed to resist blind steganalysis via embedding data in randomized locations. In addition to a concrete realization which is named the YASS algorithm in this paper, a few strategies were proposed to work with the YASS algorithm in order to enhance the data embedding rate and security. In this work, the YASS algorithm and these strategies, together referred to as YASS, have been analyzed from a warden's perspective. It is observed that the embedding locations chosen by YASS are not randomized enough and the YASS embedding scheme causes detectable artifacts. We present a steganalytic method to attack the YASS algorithm, which is facilitated by a specifically selected steganalytic observation domain (SO-domain), a term to define the domain from which steganalytic features are extracted. The proposed SO-domain is not exactly, but partially accesses, the domain where the YASS algorithm embeds data. Statistical features generated from the SO-domain have demonstrated high effectiveness in detecting the YASS algorithm and identifying some embedding parameters. In addition, we discuss how to defeat the above-mentioned strategies of YASS and demonstrate a countermeasure to a new case in which the randomness of the embedding locations is enhanced. The success of detecting YASS by the proposed method indicates a properly selected SO-domain is beneficial for steganalysis and confirms that the embedding locations are of great importance in designing a secure steganographic scheme.      

A high payload steganographic algorithm based on edge detection

The least-significant-bit (LSB) technique is one of the commonly used steganographic algorithms in the spatial domain. In most existing schemes, they didn’t carefully analyze the relationship between the image content itself. Hence, the smooth areas in the cover image will inevitably be contaminated after hiding even at a low embedding rate, thereby leading to poor visual quality and low security. In recent years, diverse steganography methods using edge detection have been proposed. However, their schemes employ certain pixels in the cover image for the sake of storing edge information, resulting in significant embedding distortion and low payload. In this study, a novel steganography approach based on the combination of LSB substitution mechanism and edge detection is proposed. To avoid the excavation of human visual system (HVS) when more secret bits are embedded into pixels, we classify the cover pixels into edge areas and non-edge areas. Then, pixels that belong to the edge area are used to carry more secret bits. In addition, to further increase the payload as well as preserve good image quality, we adopt a skillful way that the edge information is determined by most significant bits (MSBs) of the cover image so that it does not need to be stored. In the extraction phase, the same edge information is obtained. Therefore, the secret data can be correctly extracted without confusion. The experimental results demonstrate that our scheme achieves a much higher payload and better visual quality than those of state-of-the-art schemes.     

Detection of MP3Stego exploiting recompression calibration-based feature

MP3Stego is a typical steganographic tool for MP3 audio. Once the cover audio is unavailable, it is hard to distinguish between background noise and steganographic distortion. In this work, the MP3Stego algorithm has been analyzed from a warden’s perspective. It is observed that the number of bits in the bit reservoir will be disturbed when the secret message is embedded. In addition, a reliable estimation of cover audio is obtained by the proposed recompression calibration. The calibrated features are classified with support vector machine technique. Experimental results show that the proposed scheme is effective and gets good performance, especially when the embedding rate is not less than 0.01 %. The results also shows that the proposed scheme can achieve lower false positive rate comparing to the existing algorithms.     

一种基于补丁算法改进的图像隐写方法简

针对经典补丁算法进行信息隐藏时嵌入容量极低的缺陷,基于图像置乱和分块的方法,提出一种改进的图像隐写算法。首先将载体图像进行恰当分块并进行充分置乱,置乱程度以分块的均方差来衡量,然后在置乱图中选择分块对进行信息嵌入。实验结果表明,该方法可将载体图像的嵌入容量由1比特扩展到数百比特,同时具有视觉不可觉察性,对JPEG压缩和噪声攻击有较好的鲁棒性,可抵御卡方、RS和SPA分析等平面域的经典隐写分析算法。     

基于合成语音的计算安全隐写方法

计算安全的隐写理论很早就被提出,但一直不能用于主流的以多媒体数据为载体的隐写术。原因在于计算安全隐写的前提是可以获得载体的精确分布或可以按照载体分布精确采样,而自然采集的图像、音/视频无法满足这个前提条件。近几年,随着深度学习的发展,多媒体生成技术逐渐成熟且在互联网上的应用越来越普遍,生成媒体成为合理的隐写载体,隐写者可以用正常的生成媒体掩盖秘密通信,即在媒体生成过程中隐写信息,并与正常的生成媒体不可区分。一些生成模型学到的分布是可知或可控的,这将为计算安全隐写推向实用提供契机。以当前广泛应用的合成语音模型为例,设计并实现了计算安全的对称密钥隐写算法,即在音频生成过程中,根据样本点的条件概率,按算术编码的译码过程将消息解压缩到合成音频中,消息接收方拥有相同的生成模型,通过复现音频合成过程完成消息提取。在该算法的基础上进一步设计了公钥隐写算法,为实现包括隐蔽密钥交换在内的全流程隐蔽通信提供了算法支撑,在保证隐写内容安全的同时,还可以实现隐写行为安全。理论分析显示,所提隐写算法的安全性由嵌入消息的随机性决定,隐写分析实验进一步验证了当前技术下攻击者无法区分合成的载体音频与载密音频。     

基于Markov链安全性的二阶统计保持隐写算法

如何设计高阶统计安全的大容量隐写算法是当前隐写研究的难点和热点。该文基于Markov链安全性指标和动态补偿的思路,提出一种二阶统计保持的隐写算法。该算法在不降低嵌入量的前提下尽量保持了载体图像的二阶统计特性。实验结果表明,该算法在较大容量数据嵌入过程中,能较好保持二阶统计特性,取得隐写安全性的提高。     

A high stego-image quality steganographic scheme with reversibility and high payload using multiple embedding strategy

Tian’s method is a breakthrough reversible data embedding scheme with high embedding capacity measured by bits per pixel (bpp) and good visual quality measured by peak signal-to-noise ratio (PSNR). However, the embedding capacity and visual quality of this method can be significantly improved. Thus, we propose a simple reversible steganographic scheme in spatial domain for digital images by using the multiple embedding strategy. The proposed method horizontally and vertically embeds one secret bit into one cover pixel pair. The experimental results show that the proposed reversible steganographic method achieves good visual quality and high embedding capacity. Specifically, with the one-layer embedding, the proposed method can obtain the embedding capacity of more than 0.5 bpp and the PSNR value greater than 54 dB for all test images. Especially, with the five-layer embedding, the proposed method has the embedding capacity of more than 2 bpp and the PSNR value higher than 52 dB for all test images. Therefore, the proposed method surpasses many existing reversible data embedding methods in terms of visual quality and embedding capacity.     

Efficient steganalysis of images: learning is good for anticipation

This paper describes a Bayesian formalism for digital image steganalysis allowing the detection of stego images, the identification of the steganographic algorithm used, the estimation of message length and location, and anticipation in the case of embedding using an unknown steganographic algorithm. A Bayesian multinomial logistic regression based on a variational approximation is proposed. Detection, identification, and anticipation involve discriminative learning in feature space. Estimation requires the fusion of classifiers allowing discrimination between fully embedded and cover subimages of different sizes. The validation on JPEG images shows that the proposed scheme is effective and allows the anticipation of unknown steganographic algorithms.     

面向芯片实现的LTE Turbo码参数优化设计

该文分析了LTE系统中Turbo码的译码,并针对定点运算的特点进行了改进。通过大量的仿真确定了定点LOG-MAP译码算法的关键参数。量化比特数和AWGN中GAIN值,外在(extrinsic)信息因子以及定点化、归一化、溢出处理等操作,这些处理使得该译码算法在译码性能上接近浮点LOG-MAP译码算法,对于Turbo译码在硬件上的实现具有一定的参考价值。