Information Hiding
and its Applications
Steganography and Watermarking
A detailed look at Steganography
   ◦ Text Steganography
   ◦ Hypertext Steganography
   ◦ Audio Steganography
   ◦ Image Steganography
   ◦ Steganography in Open System
Image Steganography Techniques
   ◦ Spatial Domain LSB Insertion
   ◦ Masking and Filtering
   ◦ DCT-based Steganography
   ◦ Wavelet-based Steganography
How to Detect Steganography
   ◦ Blind Detection
   ◦ Analytical Detection

• Overview Papers and Articles about Steganography and Steganalysis.
• Articles and Papers about Image-based Steganography Methods.
• Theses about Steganography and Steganalysis.
• Articles and Papers about Steganalysis.

• A Mathematical Analysis of the DCT Coefficient Distributions for Images, by Edmund Y. Lam and Joseph W. Goodman, 2000.
• Steganalysis Using Image Quality Metrics, by Ismail Avcıbas, Nasir Memon and Bülent Sankur, 2003.
• Steganalysis of JPEG Images: Breaking the F5 Algorithm, by Jessica Fridrich, Miroslav Goljan, Dorin Hogea.
• Feature-Based Steganalysis for JPEG Images and Its Implications for Future Design of Steganographic Schemes, by Jessica
  Fridrich, 2004.
• Steganalysis Using Higher-Order Image Statistics, by Siwei Lyu and Hany Farid, 2006.
• Universal Detection of JPEG Steganography, by Johann Barbier, Eric Filiol, Kichenakoumar Mayoura, 2007.

A Mathematical Analysis of the DCT Coefficient Distributions for Images

By Edmund Y. Lam, Joseph W. Goodman, 2000.

ABSTRACT:

Over the past two decades, there have been various studies on the distributions of the DCT coefficients for images. However, they have concentrated only on fitting the empirical data from some standard pictures with a variety of well- known statistical distributions, and then comparing their goodness-of-fit. The Laplacian distribution is the dominant choice balancing simplicity of the model and fidelity to the empirical data. Yet, to the best of our knowledge, there has been no mathematical justification as to what gives rise to this distribution. In this paper, we offer a rigorous mathematical analysis using a doubly stochastic model of the images, which not only provides the theoretical explanations necessary, but also leads to insights about various other observations from the literature. This model also allows us to investigate how certain changes in the image statistics could affect the DCT coefficient distributions.

Steganalysis Using Image Quality Metrics

By Ismail Avcıbas, Nasir Memon and Bülent Sankur, 2003.

ABSTRACT:

We present techniques for steganalysis of images that have been potentially subjected to steganographic algorithms, both within the passive warden and active warden frameworks. Our hypothesis is that steganographic schemes leave statistical evidence that can be exploited for detection with the aid of image quality features and multivariate regression analysis. To this effect image quality metrics have been identified based on the analysis of variance (ANOVA) technique as feature sets to distinguish between cover-images and stego-images. The classifier between cover and stego-images is built using multivariate regression on the selected quality metrics and is trained based on an estimate of the original image. Simulation results with the chosen feature set and wellknown watermarking and steganographic techniques indicate that our approach is able with reasonable accuracy to distinguish between cover and stego images.

Steganalysis of JPEG Images: Breaking the F5 Algorithm

By Jessica Fridrich, Miroslav Goljan, Dorin Hogea.

ABSTRACT:

In this paper, we present a steganalytic method that can reliably detect messages (and estimate their size) hidden in JPEG images using the steganographic algorithm F5. The key element of the method is estimation of the cover-image histogram from the stego-image. This is done by decompressing the stego-image, cropping it by four pixels in both directions to remove the quantization in the frequency domain, and recompressing it using the same quality factor as the stego-image. The number of relative changes introduced by F5 is determined using the least square fit by comparing the estimated histograms of selected DCT coefficients with those of the stegoimage. Experimental results indicate that relative modifications as small as 10% of the usable DCT coefficients can be reliably detected. The method is tested on a diverse set of test images that include both raw and processed images in the JPEG and BMP formats.

Feature-Based Steganalysis for JPEG Images and Its Implications for Future Design of Steganographic Schemes

By Jessica Fridrich.

ABSTRACT:

In this paper, we introduce a new feature-based steganalytic method for JPEG images and use it as a benchmark for comparing JPEG steg-anographic algorithms and evaluating their embedding mechanisms. The detec-tion method is a linear classifier trained on feature vectors corresponding to cover and stego images. In contrast to previous blind approaches, the features are calculated as an L1 norm of the difference between a specific macroscopic functional calculated from the stego image and the same functional obtained from a decompressed, cropped, and recompressed stego image. The functionals are built from marginal and joint statistics of DCT coefficients. Because the features are calculated directly from DCT coefficients, conclusions can be drawn about the impact of embedding modifications on detectability. Three dif-ferent steganographic paradigms are tested and compared. Experimental results reveal new facts about current steganographic methods for JPEGs and new de-sign principles for more secure JPEG steganography.

Steganalysis Using Higher-Order Image Statistics

By Siwei Lyu and Hany Farid, 2006.

ABSTRACT:

Techniques for information hiding (steganography) are becoming increasingly more sophisticated and widespread. With high-resolution digital images as carriers, detecting hidden messages is also becoming considerably more difficult.We describe a universal approach to steganalysis for detecting the presence of hidden messages embedded within digital images. We show that, within multiscale, multiorientation image decompositions (e.g., wavelets), first- and higher-order magnitude and phase statistics are relatively consistent across a broad range of images, but are disturbed by the presence of embedded hidden messages.We show the efficacy of our approach on a large collection of images, and on eight different steganographic embedding algorithms.

Universal Detection of JPEG Steganography

By Johann Barbier, Eric Filiol, Kichenakoumar Mayoura, 2007.

ABSTRACT:

In this paper, we present a novel universal approach which consists in exploring statistics in the compressed frequency domain. This approach is motivated by two main characteristics of the lossless compression step of the JPEG format. First, this step can be considered as a bijective mapping and then, when only few bits are flipped at its input, half the bits are flipped at the output. These properties, combined with a binary entropy deviation we pointed out, enable the design of detection schemes which the efficiencies are constant and do not depend in practice on the amount of information that has been embedded. These characteristics define a new class of promising functions for steganalysis. We illustrate our technique by considering RLE plus Huffman as such a function and design a new efficient universal steganalytic scheme to blindly detect the use of Outguess, F5 and JPHide and JPseek. Experimental results show that our steganalysis scheme is able to efficiently detect the use of new algorithms which are not used during the training step, even if the embedding rate is very low ( 10−6). As expected, the accuracy of our detector is independent of the payload.