Detection of Diversified Stego Sources with CNNs

Butora, Jan; Fridrich, Jessica

doi:10.2352/issn.2470-1173.2019.5.mwsf-534

Cited by 11 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• The second school of thought is atomistic (= partitioned, microscopic, analytical, of divide-and-conquer type, or individualized) and consists of partitioning the distribution [73], that is to say to create a partition and to associate a classifier for each cell of the partition. Note that an example of an atomistic approach for stego-mismatch management, using a CNN multi-classifier, is presented in [11] (a class is associated with each embedding algorithm -there is thus a latent partition). Note that this idea [11], among others, has been used by the winners of the Alaska challenge [110].…”

Section: Discussion About the Mismatch Phenomenon Scenariomentioning

confidence: 99%

“…Note that an example of an atomistic approach for stego-mismatch management, using a CNN multi-classifier, is presented in [11] (a class is associated with each embedding algorithm -there is thus a latent partition). Note that this idea [11], among others, has been used by the winners of the Alaska challenge [110]. Note that again, this scenario does not consider that the test set can be used during learning.…”

Section: Discussion About the Mismatch Phenomenon Scenariomentioning

confidence: 99%

“…The idea is that she uses 3 algorithms (2 algorithms in the simplified version) that we name agents. Each of these agents will play the role of Alice, Bob 11 and Eve, and each agent runs at Alice's home. Let us note these three algorithms running at Alice's home: Agent-Alice, Agent-Bob, and Agent-Eve.…”

Section: Steganography By Deep-learningmentioning

confidence: 99%

See 2 more Smart Citations

Deep Learning in steganography and steganalysis from 2015 to 2018

Chaumont¹

2019

Preprint

View full text Add to dashboard Cite

For almost 10 years, the detection of a hidden message in an image has been mainly carried out by the computation of Rich Models (RM), followed by classification using an Ensemble Classifier (EC). In 2015, the first study using a convolutional neural network (CNN) obtained the first results of steganalysis by Deep Learning approaching the performances of the two-step approach (EC + RM). Between 2015-2018, numerous publications have shown that it is possible to obtain improved performances, notably in spatial steganalysis, JPEG steganalysis, Selection-Channel-Aware steganalysis, and in quantitative steganalysis. This chapter deals with deep learning in steganalysis from the point of view of current methods, by presenting different neural networks from the period 2015-2018, that have been evaluated with a methodology specific to the discipline of steganalysis. The chapter is not intended to repeat the basic concepts of machine learning or deep learning. So, we will present the structure of a deep neural network, in a generic way and present the networks proposed in existing literature for the different scenarios of steganalysis, and finally, we will discuss steganography by deep learning.

show abstract

Section: Discussion About the Mismatch Phenomenon Scenariomentioning

confidence: 99%

Section: Discussion About the Mismatch Phenomenon Scenariomentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning in steganography and steganalysis from 2015 to 2018

Chaumont¹

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…The classifier is a SVM with 23 features, most of which are calculated from the DCT coefficients. The work of Butora and Fridrich [5] utilizes CNNs to detect the usage of various stego embedders.…”

Section: Steganalysis Methodsmentioning

confidence: 99%

Cover-aware Steganalysis

Bunzel

Steinebach

Liu

2021

JCSANDM

View full text Add to dashboard Cite

The increasing digitization offers new ways, possibilities and needs for a secure transmission of information. Steganography and its analysis constitute an essential part of IT-Security. In this work we show how methods of blind-steganalysis can be improved to work in cover-aware scenarios, we will call this non-blind steganalysis. The main objective was to examine how to take advantage of the knowledge of reference images to maximize the accuracy-rate of the analysis. Therefore we evaluated common stego-tools and their embedding algorithms and established a dataset of 353110 images. The images have been applied to test the potency of the improved methods of the non-blind steganalysis. The results show that the accuray can be significantly improved by using cover-images to produce reference images. Also the aggregation of the outcomes has shown to have a positive impact on the accuracy. Particularly noteworthy is the correlation between the qualities of the stego- and coverimages. Only by consindering both, the accuracy could strongly be improved. Interestingly the difference between both qualities also has a deep impact on the results.

show abstract

“…Recently, considerable attention has been paid to making steganalysis more practical. It ranges from identifying diversified stego images inserted by various steganography algorithms [20], [21] to detecting hidden messages on color images of an arbitrary size with various post-processing techniques, such as the demosaicing algorithm, resizing factor, denoising, sharpening, and enhancements tools [20]- [22]. To date, SRNet has been adopted as a baseline network to extend the practicality of steganalysis for these tasks because of its training ability and performance.…”

Section: A Steganalysismentioning

confidence: 99%

Local-Source Enhanced Residual Network for Steganalysis of Digital Images

et al. 2020

View full text Add to dashboard Cite

Steganalysis refers to the study of identifying hidden messages in images inserted by steganography. Although detection performance is greatly improved when adopting convolutional neural networks (CNNs), they require sophisticated tricks, such as preprocessing for suppression of image content, using absolute and truncated activation functions, and utilizing domain knowledge. These tricks help networks train stably and mitigate the convergence problem of early stages in training, but they also restrict the flexibility of CNNs, which limits their performance. In this paper, we propose a local-source enhanced residual network (LSER) with end-to-end learning. The LSER is simple in its architecture but has two distinct characteristics from previous methods. First, the LSER uses residual blocks without any normalization. We find batch normalization is an unnecessary module in our framework. Second, a localsource skip connection is added to bypass features of different levels, which allows more abundant feature representation. Moreover, the LSER exhibits state-of-the-art results compared with the existing work in both spatial and JPEG domain steganalysis. Furthermore, a simple self-ensemble method further improves its performance without any side information. INDEX TERMSSteganalysis, low-level signal classification, convolutional neural networks, inconsistencies in noise pattern

show abstract

Detection of Diversified Stego Sources with CNNs

Cited by 11 publications

References 0 publications

Deep Learning in steganography and steganalysis from 2015 to 2018

Deep Learning in steganography and steganalysis from 2015 to 2018

Cover-aware Steganalysis

Local-Source Enhanced Residual Network for Steganalysis of Digital Images

Contact Info

Product

Resources

About