IS&T | Library

Secure Payload Scaling For Source Adaptive Payload Allocation

Abstract

In batch steganography, the sender communicates a secret message by hiding it in a bag of cover objects. The adversary performs the so-called pooled steganalysis in that she inspects the entire bag to detect the presence of secrets. This is typically realized by using a detector trained to detect secrets within a single object, applying it to all objects in the bag, and feeding the detector outputs to a pooling function to obtain the final detection statistic. This paper deals with the problem of building the pooler while keeping in mind that the Warden will need to be able to detect steganography in variable size bags carrying variable payload. We propose a flexible machine learning solution to this challenge in the form of a Transformer Encoder Pooler, which is easily trained to be agnostic to the bag size and payload and offers a better detection accuracy than previously proposed poolers.

Digital Library: EI

Published Online: February 2025

Proceedings

54 12

Batch Steganography
Deep Learning
Pooled Steganalysis
Square Root Law
Steganalysis
Steganography

Eli Dworetzky, Edgar Kaziakhmedov, Jessica Fridrich

DOI

10.2352/EI.2024.36.4.MWSF-337

Volume 36

Issue 4

Abstract

Assuming that Alice commits to an embedding method and the Warden to a detector, we study how much information Alice can communicate at a constant level of statistical detectability over potentially infinitely many uses of the stego channel. When Alice is allowed to allocate her payload across multiple cover objects, we find that certain payload allocation strategies that are informed by a steganography detector exhibit super-square root secure payload (scaling exponent 0.85) for at least tens of thousands of uses of the stego channel. We analyze our experiments with a source model of soft outputs of the detector across images and show how the model determines the scaling of the secure payload.

Digital Library: EI

Published Online: January 2024

JPEG Steganalysis Detectors Scalable With Respect to Compression Quality

78 11

Steganalysis
JPEG
Diverse source
CNN
Quality factors
Deep learning

Yassine Yousfi, Jessica Fridrich

Pages 75-1 - 75-11, January 2020, © Society for Imaging Science and Technology 2020

DOI

10.2352/ISSN.2470-1173.2020.4.MWSF-075

Volume 32

Issue 4

Practical steganalysis inevitably involves the necessity to deal with a diverse cover source. In the JPEG domain, one key element of the diversification is the JPEG quality factor, or, more generally, the JPEG quantization table used for compression. This paper investigates experimentally the scalability of various steganalysis detectors w.r.t. JPEG quality. In particular, we report that CNN detectors as well as older feature-based detectors have the capacity to contain the complexity of multiple JPEG quality factors within a single model when the quality factors are properly grouped based on their quantization tables. Detectors trained on multiple JPEG qualities show no loss of detection accuracy when compared with dedicated detectors trained for a specific JPEG quality factor. We also demonstrate that CNNs (but not so much feature-based classifiers) trained on multiple qualities can generalize to unseen custom quantization tables compared to detectors trained for specific JPEG qualities. Their ability to generalize to very different quantization tables, however, remains a challenging task. A semi-metric comparing quantization tables is introduced and used to interpret our results.

Digital Library: EI

Published Online: January 2020

Detection of Malicious Spatial-Domain Steganography over Noisy Channels Using Convolutional Neural Networks

243 13

Image steganography
Malicious steganography
Steganalysis
Transmission over noisy channels
Deep learning
Convolutional Neural Networks

Swaroop Shankar Prasad, Ofer Hadar, Ilia Polian

Pages 76-1 - 76-7, January 2020, © Society for Imaging Science and Technology 2020

DOI

10.2352/ISSN.2470-1173.2020.4.MWSF-076

Volume 32

Issue 4

Image steganography can have legitimate uses, for example, augmenting an image with a watermark for copyright reasons, but can also be utilized for malicious purposes. We investigate the detection of malicious steganography using neural networkbased classification when images are transmitted through a noisy channel. Noise makes detection harder because the classifier must not only detect perturbations in the image but also decide whether they are due to the malicious steganographic modifications or due to natural noise. Our results show that reliable detection is possible even for state-of-the-art steganographic algorithms that insert stego bits not affecting an image’s visual quality. The detection accuracy is high (above 85%) if the payload, or the amount of the steganographic content in an image, exceeds a certain threshold. At the same time, noise critically affects the steganographic information being transmitted, both through desynchronization (destruction of information which bits of the image contain steganographic information) and by flipping these bits themselves. This will force the adversary to use a redundant encoding with a substantial number of error-correction bits for reliable transmission, making detection feasible even for small payloads.

Digital Library: EI

Published Online: January 2020

Detection of Diversified Stego Sources with CNNs

74 11

Steganalysis
Steganography
Deep learning

Jan Butora, Jessica Fridrich

Pages 534-1 - 534-11, January 2019, © Society for Imaging Science and Technology 2019

DOI

10.2352/ISSN.2470-1173.2019.5.MWSF-534

Volume 31

Issue 5

The goal of this article is construction of steganalyzers capable of detecting a variety of embedding algorithms and possibly identifying the steganographic method. Since deep learning today can achieve markedly better performance than other machine learning tools, our detectors are deep residual convolutional neural networks. We explore binary classifiers trained as cover versus all stego, multi-class detectors, and bucket detectors in a feature space obtained as a concatenation of features extracted by networks trained on individual stego algorithms. The accuracy of the detector to identify steganography is compared with dedicated detectors trained for a specific embedding algorithm. While the loss of detection accuracy w.r.t. increasing number of steganographic algorithms increases only slightly as long as the embedding schemes are known, the ability of the detector to generalize to previously unseen steganography remains a challenging task.

Digital Library: EI

Published Online: January 2019

Algorithm Mismatch in Spatial Steganalysis

43 9

Steganography
Steganalysis
Domain adaptation
Blind detection

Stephanie Reinders, Li Lin, Yong Guan, Min Wu, Jennifer Newman

Pages 535-1 - 535-11, January 2019, © Society for Imaging Science and Technology 2019

DOI

10.2352/ISSN.2470-1173.2019.5.MWSF-535

Volume 31

Issue 5

The number and availability of stegonographic embedding algorithms continues to grow. Many traditional blind steganalysis frameworks require training examples from every embedding algorithm, but collecting, storing and processing representative examples of each algorithm can quickly become untenable. Our motivation for this paper is to create a straight-forward, nondata-intensive framework for blind steganalysis that only requires examples of cover images and a single embedding algorithm for training. Our blind steganalysis framework addresses the case of algorithm mismatch, where a classifier is trained on one algorithm and tested on another, with four spatial embedding algorithms: LSB matching, MiPOD, S-UNIWARD and WOW. We use RAW image data from the BOSSbase database and and data collected from six iPhone devices. Ensemble Classifiers with Spatial Rich Model features are trained on a single embedding algorithm and tested on each of the four algorithms. Classifiers trained on MiPOD, S-UNIWARD and WOW data achieve decent error rates when testing on all four algorithms. Most notably, an Ensemble Classifier with an adjusted decision threshold trained on LSB matching data achieves decent detection results on MiPOD, S-UNIWARD and WOW data.

Digital Library: EI

Published Online: January 2019

StegoAppDB: a Steganography Apps Forensics Image Database

91 11

Steganography
Steganalysis
Database
Mobile phone camera
Image forensics

Jennifer Newman, Li Lin, Wenhao Chen, Stephanie Reinders, Yangxiao Wang, Min Wu, Yong Guan

DOI

10.2352/ISSN.2470-1173.2019.5.MWSF-536

Volume 31

Issue 5

In this paper, we present a new reference dataset simulating digital evidence for image (photographic) steganography. Steganography detection is a digital image forensic topic that is relatively unknown in practical forensics, although stego app use in the wild is on the rise. This paper introduces the first database consisting of mobile phone photographs and stego images produced from mobile stego apps, including a rich set of side information, offering simulated digital evidence. StegoAppDB, a steganography apps forensics image database, contains over 810,000 innocent and stego images using a minimum of 10 different phone models from 24 distinct devices, with detailed provenanced data comprising a wide range of ISO and exposure settings, EXIF data, message information, embedding rates, etc. We develop a camera app, Cameraw, specifically for data acquisition, with multiple images per scene, saving simultaneously in both DNG and high-quality JPEG formats. Stego images are created from these original images using selected mobile stego apps through a careful process of reverse engineering. StegoAppDB contains cover-stego image pairs including for apps that resize the stego dimensions. We retain the original devices and continue to enlarge the database, and encourage the image forensics community to use StegoAppDB. While designed for steganography, we discuss uses of this publicly available database to other digital image forensic topics.

Digital Library: EI

Published Online: January 2019

Steganalyzing Images of Arbitrary Size with CNNs

244 38

Steganalysis
Deep Learning
Size Independent

Clement Fuji Tsang, Jessica Fridrich

DOI

10.2352/ISSN.2470-1173.2018.07.MWSF-121

Volume 30

Issue 7

Convolutional neural networks offer much more accurate detection of steganography than the outgoing paradigm - classifiers trained on rich representations of images. While training a CNN is scalable with respect to the size of the training set, one cannot directly train on images that are too large due to the memory limitations of current GPUs. Most leading network architectures for steganalysis today require the input image to be a small tile with 256 × 256 or 512 × 512 pixels. Because detecting the presence of steganographic embedding changes really means detecting a very weak noise signal added to the cover image, resizing an image before presenting it to a CNN would be highly suboptimal. Applying the tile detector on disjoint segments of a larger image and fusing the results bring a plethora of new problems of how to properly fuse the outputs. In this paper, we propose a different solution to this problem based on modifying an existing leading network architecture for steganalysis in the spatial domain, the YeNet, to output statistical moments of feature maps to the fully-connected classifier part of the network. On experiments in which we adjust the payload with image size according the square root law for constant statistical detectability, we demonstrate that the proposed architecture can be trained to steganalyze images of various sizes without any or only a small loss with respect to detectors trained for a fixed image size.

Digital Library: EI

Published Online: January 2018

Natural Steganography in JPEG Compressed Images

51 2

Steganography
Steganalysis
Natural steganography
JPEG

Tomáš Denemark, Patrick Bas, Jessica Fridrich

DOI

10.2352/ISSN.2470-1173.2018.07.MWSF-316

Volume 30

Issue 7

In natural steganography, the secret message is hidden by adding to the cover image a noise signal that mimics the heteroscedastic noise introduced naturally during acquisition. The method requires the cover image to be available in its RAW form (the sensor capture). To bring this idea closer to a practical embedding method, in this paper we embed the message in quantized DCT coefficients of a JPEG file by adding independent realizations of the heteroscedastic noise to pixels to make the embedding resemble the same cover image acquired at a larger sensor ISO setting (the so-called cover source switch). To demonstrate the feasibility and practicality of the proposed method and to validate our simplifying assumptions, we work with two digital cameras, one using a monochrome sensor and a second one equipped with a color sensor. We then explore several versions of the embedding algorithm depending on the model of the added noise in the DCT domain and the possible use of demosaicking to convert the raw image values. These experiments indicate that the demosaicking step has a significant impact on statistical detectability for high JPEG quality factors when making independent embedding changes to DCT coefficients. Additionally, for monochrome sensors or low JPEG quality factors very large payload can be embedded with high empirical security.

Digital Library: EI

Published Online: January 2018

How to augment a small learning set for improving the performances of a CNN-based steganalyzer?

39 1

Steganalysis
Deep Learning
Convolutional Neural Network
Base augmentation

Mehdi Yedroudj, Marc Chaumont, Frédéric Comby

DOI

10.2352/ISSN.2470-1173.2018.07.MWSF-317

Volume 30

Issue 7