Abstract.A new method of text steganography based on Markov chains of different orders that allows the introduction of hidden information in texts is presented together with test results of a software solution which generate texts with a good approximation to the natural language model. Keywords: text steganography, Markov process, automatic text generation, text naturalness. INTRODUCTIONSteganography is a science of a hidden data transmission by the concealment of the fact of data transfer. Currently, the problem of steganographic data protection from unauthorized access is extremely urgent [1]. However, most studies in this area are focused either on a concept of hidden information embedded in multimedia containers (images, audio and video files) of different formats, or aimed at the use of telecommunication networks (network steganography). At the same time, the development of linguistic steganography, which uses text information as a container, has received too little attention. This is explained by the fact that the steganographic methods based on the embedding of hidden information in media files, as well as in telecommunication networks, in fact, are not suitable when using natural language texts as a steganographic container. Nevertheless, methods of text steganography can be widely used, since, according to statistics, it is textual information that has the highest transmission intensity [2]. The advantage of text containers over other media containers lies in the fact that methods of analysis of text files for the presence of hidden information have not been fully implemented at present [3]. At the same time, there are many algorithms designed exclusively for text containers, a description of which can be found, for example, in [4,5]. Currently, there are a lot of methods of text steganography based on the use of Markov models. For example, in [2, 6] input data is used for text generation using Markov chains. However, the proposed models are greatly simplified in order to facilitate calculation, since it is assumed that all probabilities of transition from a given state to any other are equal. The submitted paper is based on the studies of [7], in which the probabilities of transition of one word to another are maintained with sufficiently fair accuracy. The novelty of the method presented in this paper is that the proposed steganographic method is based on higher-order Markov processes (second and third) and also allows for working with Russian texts.

The paper proposes a method of extracting the feature vector of images, which makes it possible to effectively detect the presence of hidden information in JPEG images embedded by various popular steganography tools. This method is based on the usage of the transition probability matrix. The essence of the method for extracting the feature vector of the image is to use the transition probability matrix and apply the image calibration method to improve the accuracy of steganalysis and reduce the number of false positives. For each image from the training and test sets a feature vector is found in this way, the number of elements is 324. Further, the models were trained on the training dataset by each of machine learning methods separately: decision trees with gradient boosting, linear models, k-nearest neighbors, support vector machines, neural networks, and artificial immune systems. To assess the capacity of the models the following metrics were used: accuracy, the rate of the false positive and false negative errors, and the confusion matrix. The results of classification by each of the above methods are given. For training and testing a dataset IStego100K was used, which consists of 208 thousand images of the same size 1024 x 1024 with different quality values in the range from 75 to 95. One of the J-UNIWARD, nsF5, and UERD steganography algorithms was used to embed a hidden message. As a result, we can observe that the proposed approach to extracting the feature vector makes it possible to detect the presence of hidden information embedded by non-adaptive steganography (Steghide, OutGuess and nsF5) in static JPEG images with high accuracy (more than 95%). However, for adaptive steganography methods (J-UNIWARD, UERD) the accuracy is less (about 50-60%).

The purpose of this work is to develop the steganalysis method of static JPEG images, based on the usage of artificial immune systems.A model of an artificial immune system was developed for the problem of detecting hidden information in JPEG images. Basic requirements were determined, and basic elements of an artificial immune system were considered, mutation and antibody cloning operations were introduced. Also, formal description of main nodes of the artificial immune system is presented. In addition, a brief overview and analysis of the state of the steganalysis problem is provided in the paper. Furthermore, analysis of the obtained experimental results and an assessment of the effectiveness are performed for the developed method.The proposed method allows to detect the presence of hidden information, embedded by various popular steganography tools (like OutGuess, Steghide and F5) in static JPEG images with a sufficiently high accuracy. The theoretical significance of this work consists in the development of a fairly promising approach of heuristic steganalysis using artificial immune systems. The practical significance lies in the developed software product, as well as in experimental data, that confirms the effectiveness of the steganalysis method towards the detection of hidden information in JPEG images.

No abstract

Рассмотрены современные методы организации скрытых сетевых каналов передачи информации. Выдвинуто предположение об эффективности использования протоколов потоковой передачи данных для организации скрытых каналов. Предложен метод скрытого обмена информацией в открытых сетях. Приведена функциональная модель стеганосистемы на основе протокола RTP и показан ее программный прототип. Приведены оценочные характеристики стеганосистемы. Показаны результаты эксплуатационного тестирования программного прототипа стеганосистемы в лабораторных условиях и в сети Интернет. Программный прототип показал высокую скрытность при приемлемой для многих задач пропускной способности. Вместе с тем выявлено снижение передающих характеристик системы по мере усложнения маршрутов передачи сетевой среды. Полученные результаты исследований имеют две важнейшие области применения. Методики детектирования нелегальных сетевых стеганоканалов могут быть использованы разработчиками DLP-систем, правоохранительными органами и оборонными ведомствами. Предлагаемый метод скрытой передачи информации может быть использован для организации телеметрического канала системы связи, например, спутниковой. Ключевые слова стеганография, стеганосистема, скрытый информационный канал, скрытая передача информации, сетевая стеганография, протокол потоковой передачи данных