Search

The paper proposes a modification of the steganographic method of embedding a digital watermark into Wavelet transform domain of a digital image. The proposed method embeds a watermark in the high-frequency detail coefficients of a three-level Wavelet transform using the Dobeshi-43 wavelet, which ensures the high quality of the resulting stego. Stego is a result of embedding of digital watermark into container and is saved in a losses format that is caused by wide dissemination of such images due to their small size. Discrete cosine transform is applied to a digital watermark, thereby increasing the efficiency of detecting the presence of a watermark in a stego. In work the analysis of efficiency of detecting of the digital watermark depending on the threshold values defining the choice of the detailing coefficients used for embedding is carried out. A result of this analysis is establishment the fact that increase in thresholds without application of Discrete cosine transformation to the watermark leads to considerable deterioration in accuracy of detecting. Conversely, the proposed method with the threshold values Т1=80 and Т2=90 of the choice of detailing coefficients in conjunction with the Discrete Cosine Transformation of the watermark matrix increased the percentage of correctly identified stegos. The results of the experiments showed the resistance of embedding a digital watermark to compression attacks (even with image quality QF=5), imposing noise, sharpening, shifting, for which the detection accuracy averages 97%. Also, the method is resistant to scaling provided QF>50. This paper compares the resistance of the original method of embedding a digital watermark and different variations of the proposed modification of the steganographic method to complex attacks (compression in combination with other types of digital image processing). [ABSTRACT FROM AUTHOR]

In this paper we study the effect of perturbing influences aimed at the spatial domain of digital images on the change of color triplets in the matrix of unique colors of a digital image. In the course of computational experiments established that the modification of only blue color matrix of the image, the quantity of Blue-triads (which indicate the blue color component of the image) in the matrix of unique colors of the digital image significantly exceeds the quantity of Red- and Green-triads that will allow for the examination of images clearly establish the fact of modifying the blue color component. The identified feature is typical for most containers in a lossy format which is associated with the digital image's formation technologies and compression algorithms therefore despite the low sensitivity of the human eye to blue color it is not recommended to use the blue color component for spatial transformations of the image, in particular, for steganographic transformations. During the experiments the most suitable for spatial transformations green and red color components revealed, since their modifications lead to the appearance of a smaller quantity of Green- and Red-triads (which indicate the green and red color components of the image, respectively) and to obtain a comparable quantity of all types of triads of triplets in a matrix of unique colors which will complicate the correct examination of a digital image from the point of view of unambiguously identifying a modified color component. The obtained results are substantiated from the standpoint of color perception by the human eye, which underlies the formation of colors by digital cameras, and the image compression features of the JPEG standard. Criteria for determining the color component of a digital image providing less disturbance in the quantity of sequential triads in a matrix of unique colors and the basic steps for determining for a particular image the most and least suitable for steganographic transformation color components in terms of changes in color triplets in a matrix of unique colors are given. [ABSTRACT FROM AUTHOR]

Modern information protection issues are changing with the development of information technology. Every year, the level of cybercrime is getting higher everywhere. Often cybercriminals are exposed to digital files of various formats. One of the main tasks of modern information security is the detection of violations of the integrity of digital signals, in particular, digital imaging. Detection of digital image processing is currently a topical task due to the increased incidence of integrity of important documents. Digital images are almost always processed by the filters of graphic editors. One of the popular filters is the sharpening. In the open edition, a single algorithm for detecting artificial sharpening was found. The existing algorithm is based on the analysis of close pairs of colors of a digital image in relation to the total number of pairs of colors of the matrix of a digital image. Threshold value is detected, considering what certain ratio of these indicators indicates presence or absence of sharpening. However, his main disadvantage is a large number of errors of the first and second kind. In this paper, the influence of the partition of the image matrix on square blocks of different sizes on the effectiveness of detecting sharpening is investigated. As a sharpening filter, Adobe Photoshop's Intelligent Sharpness tool is chosen because of its wide-ranging popularity. On the basis of the conducted research modification of the algorithm was performed. This modification made it possible to reduce the number of errors of the 1st kind by 12%, and the 2 genera by 4%, which results in 40% of errors of the second kind and 14% of errors of the 1st genus. Also, possible further variants of development and improvement of the process of detection of the indicated processing, namely, the study of the influence of the modification of the formula of close pairs of colors on the efficiency of the algorithm, are proposed in the work. [ABSTRACT FROM AUTHOR]

In the paper we study the effect of modification of DCT coefficients on the spatial domain of digital images using analysis of sequential triads of color triplets in a matrix of unique colors of the image, for further steganalysis of stego formed by modifying the DCT domain. On the basis of the computational experiments the improvement of the previously developed steganalytic method for detecting the presence of additional information attachments in digital images was carried out. A software interface for steganalysis of digital images was developed in MATLAB environment. [ABSTRACT FROM AUTHOR]