Merging Markov and DCT features for multi-class JPEG steganalysis

Pevný, Tomáš; Fridrich, Jessica

doi:10.1117/12.696774

Cited by 368 publications

(325 citation statements)

References 20 publications

Supporting

Mentioning

308

Contrasting

Unclassified

Order By: Relevance

“…The calibration method typically used for JPEG steganography is quite simple; a few pixel rows and/or columns are cropped from the image so as to desynchronize it from the original JPEG grid and the resulting image is compressed again, which forms a good approximation of the cover image. The results reported in [10], the most recent multi-class JPEG steganalysis method that employs such self-calibration, are close to perfect: the steganalyst can determine one out of 6 stego algorithms employed for hiding with a detection accuracy of more than 95% in most cases, even at low embedding rates.…”

Section: Introductionmentioning

confidence: 80%

“…In spite of the absence of good universal models, recent steganalysis algorithms have been very successful by using a self-calibration method to approximate the statistics of the original cover (see, for example, Pevny and Fridrich [9,10], and Dabeer et al [17]). The calibration method typically used for JPEG steganography is quite simple; a few pixel rows and/or columns are cropped from the image so as to desynchronize it from the original JPEG grid and the resulting image is compressed again, which forms a good approximation of the cover image.…”

Section: Introductionmentioning

confidence: 99%

“…It is known that, the higher order statistics, in general, are difficult to match, model, or restore. Recently, blind steganalysis algorithms [9][10][11][12][13][14][15][16] have been proposed that employ supervised learning to distinguish between the plain cover and stego images, and also identify the particular hiding algorithm used for steganography. These techniques bank on the fact that there are some image features that are modified during the embedding process which can be used as an input to the learning machine.…”

Section: Introductionmentioning

confidence: 99%

“…Many steganalysis schemes (see [8][9][10]) have been able to successfully detect the above steganographic techniques that match marginal statistics or models. They exploit the fact that higher order statistics get modified by data hiding using these stego methods.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

YASS: Yet Another Steganographic Scheme That Resists Blind Steganalysis

2007

View full text Add to dashboard Cite

Abstract.A new, simple, approach for active steganography is proposed in this paper that can successfully resist recent blind steganalysis methods, in addition to surviving distortion constrained attacks. We present Yet Another Steganographic Scheme (YASS), a method based on embedding data in randomized locations so as to disable the selfcalibration process (such as, by cropping a few pixel rows and/or columns to estimate the cover image features) popularly used by blind steganalysis schemes. The errors induced in the embedded data due to the fact that the stego signal must be advertised in a specific format such as JPEG, are dealt with by the use of erasure and error correcting codes. For the presented JPEG steganograhic scheme, it is shown that the detection rates of recent blind steganalysis schemes are close to random guessing, thus confirming the practical applicability of the proposed technique. We also note that the presented steganography framework, of hiding in randomized locations and using a coding framework to deal with errors, is quite simple yet very generalizable.

show abstract

Section: Introductionmentioning

confidence: 80%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

YASS: Yet Another Steganographic Scheme That Resists Blind Steganalysis

2007

View full text Add to dashboard Cite

show abstract

“…In the frequency domain, Pevny and Jessica [113] developed a multi-class JPEG steganalysis system that comprised of DCT features and calibrated Markov features, which were then merged to produce a 274-dimensional feature vector. This vector is fed into a Support Vector Machine multi-classifier capable of detecting the presence of Model-Based steganography, F5, OutGuess, Steghide and JP Hide&Seek.…”

Section: Fig 29mentioning

confidence: 99%

Digital image steganography: Survey and analysis of current methods

et al. 2010

View full text Add to dashboard Cite

Steganography is the science that involves communicating secret data in an appropriate multimedia carrier, e.g., image, audio, and video files. It comes under the assumption that if the feature is visible, the point of attack is evident, thus the goal here is always to conceal the very existence of the embedded data. Steganography has various useful applications. However, like any other science it can be used for ill intentions. It has been propelled to the forefront of current security techniques by the remarkable growth in computational power, the increase in security awareness by, e.g., individuals, groups, agencies, government and through intellectual pursuit. Steganography's ultimate objectives, which are undetectability, robustness (resistance to various image processing methods and compression) and capacity of the hidden data, are the main factors that separate it from related techniques such as watermarking and cryptography. This paper provides a state-of-the-art review and analysis of the different existing methods of steganography along with some common standards and guidelines drawn from the literature. This paper concludes with some recommendations and advocates for the object-oriented embedding mechanism. Steganalysis, which is the science of attacking steganography, is not the focus of this survey but nonetheless will be briefly discussed.

show abstract

HYSA: HYbrid steganographic approach using multiple steganography methods

Sajedi¹,

Jamzad²

2010

Security Comm Networks

View full text Add to dashboard Cite

The aim of image steganography is to hide data in an image with visual and statistical invisibility. There have been a number of steganography techniques proposed over the past few years. In sequence, the development of these methods has led to an increased interest in steganalysis techniques. Most of the steganalysis methods attempt to estimate cover image statistics. One way to provide a secure steganography method is to disturb the estimation of steganalyzers. The main concern of this paper is to resist against steganalysis methods and utilize a mechanism to securely embed more secret data into an image. We present HYbrid steganographic approach (HYSA), which embeds secret data with randomized embedding algorithms in randomized regions of the cover image. HYSA embeds secret data in only some regions of the image not all over the image. In addition, HYSA can employ more than one steganography method for embedding secret data in the image. Random selection of embedding regions and steganography methods leaves a combination of several types of distortion on the image, which are difficult to be recognized by the steganalysis methods. Experimental results by applying some well-known and efficient steganalysis methods illustrate that the proposed steganographic approach is more undetectable than some popular steganography methods.

show abstract

Merging Markov and DCT features for multi-class JPEG steganalysis

Cited by 368 publications

References 20 publications

YASS: Yet Another Steganographic Scheme That Resists Blind Steganalysis

YASS: Yet Another Steganographic Scheme That Resists Blind Steganalysis

Digital image steganography: Survey and analysis of current methods

HYSA: HYbrid steganographic approach using multiple steganography methods

Contact Info

Product

Resources

About