Reversible steganography based on side match and hit pattern for VQ-compressed images

Data hiding is an important technique in multimedia security. Multimedia data hiding techniques enable message senders to disguise secret data by embedding them into cover media. Thus, delivering secret messages is as easy as sending the cover media. Recently, many researchers have studied reversible data hiding for images. Those methods can reconstruct the original cover image and extract the embedded secret data from a stego-image. This study proposes a novel reversible steganographic method of embedding secret data into a vector quantization (VQ) compressed image by applying the concept of side match. The proposed method uses extra information, namely the hit pattern, to achieve reversibility. Moreover, its small hit pattern enables the embedding of the entire hit pattern along with the secret data in most cases. To optimize visual quality of the output stego-image, the method applies the concept of partitioned codebooks (state codebooks). The partition operation on the codebook uses a look-up table to minimize embedding and extraction time. We also propose the use of diagonal seed blocks to embed the entire hit pattern into the cover image without producing any extra control messages. Compared to the Chang and Lin method, the experimental results show that the proposed method has higher capacity, better visual quality, and shorter execution time.     

Listega: list-based steganography methodology

The use of textual list of items, e.g., products, subjects, books, etc., is widely popular and linguistically legible. This motivates the development of List-Based Steganography Methodology (Listega). Listega takes advantage of such use of textual list to camouflage data by exploiting itemized data to conceal messages. Simply, it encodes a message then assigns it to legitimate items in order to generate a text-cover in a form of list. The generated list of items, the text-cover, can be embedded among other legitimate noncoded items for more protection based on a predetermined protocol among communicating parties such as read every other item, every fifth item, or any other way than the use of particular sequence. Listega neither hides data in a noise (errors) nor produces noise. Instead, it camouflages data by manipulating noiseless list of legitimate items. Listega establishes a covert channel among communicating parties by employing justifiably reasons based on the common practice of using textual list of items in order to achieve unsuspicious transmission of generated covers. The presented implementation, validation, and steganalysis of Listega demonstrate: the robustness capabilities of achieving the steganographic goal, the adequate room for concealing data, and the superior bitrate of roughly 1.32 up to 3.87% than contemporary linguistic steganography approaches.     

An evaluation of image based steganography methods

Steganography is a process that involves hiding a message in an appropriate carrier for example an image or an audio file. The carrier can then be sent to a receiver without anyone else knowing that it contains a hidden message. This is a process, which can be used for example by civil rights organisations in repressive states to communicate their message to the outside world without their own government being aware of it. Less virtuously it can be used by terrorists to communicate with one another without anyone else’s knowledge. In both cases the objective is not to make it difficult to read the message as cryptography does, it is to hide the existence of the message in the first place possibly to protect the courier. The initial aim of this study was to investigate steganography and how it is implemented. Based on this work a number of common methods of steganography could then be implemented and evaluated. The strengths and weaknesses of the chosen methods can then be analysed. To provide a common frame of reference all of the steganography methods implemented and analysed used GIF images. Seven steganography methods were implemented. The methods were chosen for their different strengths in terms of resistance to different types of steganalysis or their ability to maximise the size of the message they could store. All of the methods used were based on the manipulation of the least significant bits of pixel values or the rearrangement of colours to create least significant bit or parity patterns, which correspond to the message being hidden.An erratum to this article can be found at     

Adaptive color quantization based on perceptive edge protection

A novel adaptive color quantization algorithm is presented based on perceptive information carried by image edges. A few colors can be selected automatically to protect the semantic regions and represent the original image. Experimental results show that the proposed methods perform well in term of perception and speed.     

Increasing image compression rate using steganography

The goal of image compression is to remove the redundancies for minimizing the number of bits required to represent an image while steganography works by embedding the secret data in redundancies of the image in invisibility manner. Our focus in this paper is the improvement of image compression through steganography. Even if the purposes of digital steganography and data compression are by definition contradictory, we use these techniques jointly to compress an image. Hence, two schemes exploring this idea are suggested. The first scheme combines a steganographic algorithm with the baseline DCT-based JPEG, while the second one uses this steganographic algorithm with the DWT-based JPEG. In this study data compression is performed twice. First, we take advantage of energy compaction using JPEG to reduce redundant data. Second, we embed some bit blocks within its subsequent blocks of the same image with steganography. The embedded bits not only increase file size of the compressed image, but also decrease the file size further more. Experimental results show for this promising technique to have wide potential in image coding.     

应用协同进化的图像矢量量化码书设计方法

为了减小LBG算法对初始码书的依赖性,提高跳出局部最优的能力,提出了一种基于协同进化的矢量量化码书设计方法（Coevolution Based LBG,CLBG）。该算法根据码书在同其他码书竞争中的表现来衡量码书的适应度。实验结果表明：CLBG有效地减小了算法对初始码书的依赖性,所得码书性能超过了其他典型的改进码书设计方法。     

自适应量化权重用于通信高效联邦学习

针对联邦学习训练过程中通信资源有限的问题,本文提出了两种联邦学习算法:自适应量化权重算法和权重复用控制算法,前者对权重的位数进行压缩,减少通信过程中传输的比特数,算法在迭代过程中,自适应调整量化因子,不断减少量化误差;后者能阻止不必要的更新上传,从而减少上传的比特数.基于标准检测数据集Mnist和Cifar10,在CNN和MLP网络模型上做了仿真模拟,实验结果表明,与典型的联邦平均算法相比,提出的算法降低了75%以上的通信成本.     

基于深度残差网络的迭代量化哈希图像检索方法

针对现有的哈希图像检索方法表达能力较弱、训练速度慢、检索精度低,难以适应大规模图像检索的问题,提出了一种基于深度残差网络的迭代量化哈希图像检索方法（DRITQH）。首先,使用深度残差网络对图像数据进行多次非线性变换,从而提取图像数据的特征,并获得具有语义特征的高维特征向量;然后,使用主成分分析（PCA）对高维图像特征进行降维,同时运用迭代量化对生成的特征向量进行二值化处理,更新旋转矩阵,将数据映射到零中心二进制超立方体,从而最小化量化误差并得到最佳的投影矩阵;最后,进行哈希学习,以得到最优的二进制哈希码在汉明空间中进行图像检索。实验结果表明,DRITQH在NUS-WIDE数据集上,对4种哈希码的检索精度分别为0.789、0.831、0.838和0.846,与改进深度哈希网络（IDHN）相比分别提升了0.5、3.8、3.7和4.2个百分点,平均编码时间小了1 717 μs。DRITQH在大规模图像检索时减少了量化误差带来的影响,提高了训练速度,实现了更高的检索性能。     

Varying radix numeral system based adaptive image steganography

A novel image steganographic scheme based on the primitive and renowned numerical model is proposed. In this adaptive scheme the data to be embedded is dissected into numerals, each having variable information carrying capacity. The dissection is based on the statistics of the host image pixels i.e. the amount of adulteration that a pixel can tolerate. The proposed method provides proficient visual quality despite high payload capacity. The experimental results evaluated on 300 natural images show that the new scheme is resistant to RS steganalysis and offers high visual quality when compared with typical LSB-based schemes as well as their edge adaptive versions, like pixel-value-differencing-based schemes.     

Multi-party covert communication with steganography and quantum secret sharing

In this paper, we address the “multi-party covert communication”, a stronger notion of security than standard secure multi-party communication. Multi-party covert communication guarantees that the process of it cannot be observed. We propose a scheme for steganographic communication based on a channel hidden within quantum secret sharing (QSS). According to our knowledge nobody has ever raised the scheme, providing us the motivation for this work. To an outside observer, participants will engage in a typical instance of QSS, just like the others. But when the session is over, covert multi-party communication has already been done. Further analysis shows that the amount of hidden information one can acquire is 0, even if either an outside observer guesses the covert communication is carrying on or a dishonest participant is eavesdropping.     

A framework for avoiding steganography usage over HTTP

Steganographic techniques allow users to covertly transmit information, hiding the existence of the communication itself. These can be used in several scenarios ranging from evading censorship to discreetly extracting sensitive information from an organization. In this paper, we consider the problem of using steganography through a widely used network protocol (i.e. HTTP). We analyze the steganographic possibilities of HTTP, and propose an active warden model to hinder the usage of covert communication channels. Our framework is meant to be useful in many scenarios. It could be employed to ensure that malicious insiders are not able to use steganography to leak information outside an organization. Furthermore, our model could be used by web servers administrators to ensure that their services are not being abused, for example, as anonymous steganographic mailboxes. Our experiments show that steganographic contents can be successfully eliminated, but that dealing with high payload carriers such as large images may introduce notable delays in the communication process.     

一种可抵御RS统计分析的图像信息隐藏方案

通过动态补偿隐藏信息后的图像,给出一种可抵御RS分析的LSB隐藏方案．该方案基于混沌系统产生随机序列选择嵌入位置并进行LSB嵌入,对嵌入信息后的图像进行动态补偿,补偿后的图像即使信息嵌入率接近于100％,RS方法得到的估计值仍将低于判决阈值,从而作出错误判决．混沌系统初值和动态补偿参数可进一步提高系统的安全性．实验表明,该方案可有效抵抗经典RS方法及其多种改进方法的分析．     

Classified self-organizing map with adaptive subcodebook for edge preserving vector quantization

This paper presents a novel classified self-organizing map method for edge preserving quantization of images using an adaptive subcodebook and weighted learning rate. The subcodebook sizes of two classes are automatically adjusted in training iterations based on modified partial distortions that can be estimated incrementally. The proposed weighted learning rate updates the neuron efficiently no matter of how large the weighting factor is. Experimental results show that the new method achieves better quality of reconstructed edge blocks and more spread out codebook and incurs a significantly less computational cost as compared to the competing methods.     

Adaptive compensation control for special actuator circumstances with input quantization

In this article, considering actuator constraints and possible failures, an adaptive compensation control scheme is developed to realize tracking control for a class of uncertain nonlinear systems with quantized inputs. A new variable is generated to evaluate the effect of actuator saturation and is used in the process of controller design to compensate for the influence of actuator saturation constraint. Moreover, the controller is able to show certain accommodation capability to tolerate possible actuator failures and input quantization error via integrating parameter update process of unknown fault constants into adaption of parametric uncertainties under the backstepping procedure. Specifically, actuator saturation effect and possible actuator failures as well as input quantization error can be dealt with uniformly under the framework of the proposed scheme and the control system has certain robustness to external disturbances. It is proved that all the signals of the closed‐loop system are ensured to be bounded and the tracking error is enabled to converge toward a compact set, which is adjustable by tuning design parameters. Finally, experiments are carried out on an active suspension plant to illustrate the effectiveness of the proposed control scheme.     

A novel adaptive steganography based on local complexity and human vision sensitivity

This paper presents a novel adaptive steganographic scheme that is capable of both preventing visual degradation and providing a large embedding capacity. The embedding capacity of each pixel is dynamically determined by the local complexity of the cover image, allowing us to maintain good visual quality as well as embedding a large amount of secret messages. We classify pixels into three levels based on the variance of the local complexity of the cover image. When determining which level of local complexity a pixel should belong to, we take human vision sensitivity into consideration. This ensures that the visual artifacts appeared in the stego image are imperceptible, and the difference between the original and stego image is indistinguishable by the human visual system. The pixel classification assures that the embedding capacity offered by a cover image is bounded by the embedding capacity imposed on three levels that are distinguished by two boundary thresholds values. This allows us to derive a combination ratio of the maximal embedding capacity encountered with at each level. Consequently, our scheme is capable of determining two threshold values according to the desired demand of the embedding capacity requested by the user. Experimental results demonstrated that our adaptive steganographic algorithm produces insignificant visual distortion due to the hidden message. It provides high embedding capacity superior to that offered by a number of existing schemes. Our algorithm can resist the RS steganalysis attack, and it is statistically invisible for the attack of histogram comparison. The proposed scheme is simple, efficient and feasible for adaptive steganographic applications.     

一种自适应的奇偶量化水印算法

数字水印是数字媒体版权保护的一种有效方法。提出了一种自适应的奇偶量化水印算法,其量化步长由Watson模型和JPEG量化表共同确定,并通过二者自适应地调节量化步长。水印在DCT的中频系数中嵌入,由于采用奇偶量化,水印提取方法简单。实验结果表明,该算法具有良好的不可见性,算法简单,并能抵抗常规的图像处理和噪声攻击,具有很强的鲁棒性。     

A high quality image sharing with steganography and adaptive authentication scheme

With the rapid growth of numerous multimedia applications and communications through Internet, secret image sharing has been becoming a key technology for digital images in secured storage and confidential transmission. However, the stego-images are obtained by directly replacing the least-significant-bit planes (LSB) of cover-images with secret data and authentication code in most schemes, which will result in the distortion of the stego-images. In this paper, we proposed a novel secret image sharing scheme by applying optimal pixel adjustment process to enhance the image quality under different payload capacity and various authentication bits conditions. The experimental results showed that the proposed scheme has improved the image quality of stego images by 4.71%, 9.29%, and 11.10%, as compared with the schemes recently proposed by Yang et al., Chang et al., and Lin and Tsai. We also provide several experiments to demonstrate the efficacy of authentication capability of the proposed scheme. In other words, our scheme maintains the secret image sharing and authentication ability while enhances the image quality.     

Colour quantization with Ant-tree

Ant-tree is a clustering algorithm inspired by biological ants. This paper defines a variant of such algorithm to perform colour quantization. Some of the features of the basic Ant-tree have been adapted to obtain a quicker algorithm and to perform the main steps of colour quantization on a big input set. The centroid of every cluster defines a colour of the palette and once the complete colour palette is established, the algorithm represents each pixel of the original image by the colour associated to its cluster.Computational results show that the error obtained for the quantized images is smaller than the error generated by some other well-known quantization methods.