Audio Steganography Based on Iterative Adversarial Attacks Against Convolutional Neural Networks

Recently, convolutional neural networks (CNNs) have demonstrated superior performance on digital multimedia steganalysis. However, some studies have noted that most CNN-based classifiers can be easily fooled by adversarial examples, which form slightly perturbed inputs to a target network according to the gradients. Inspired by this phenomenon, we first introduce a novel steganography method based on adversarial examples for digital audio in the time domain. Unlike related methods for image steganography, such as <xref ref-type="bibr" rid="ref1">[1]</xref>–<xref ref-type="bibr" rid="ref2"/><xref ref-type="bibr" rid="ref3"/><xref ref-type="bibr" rid="ref4">[4]</xref>, which are highly dependent on some existing embedding costs, the proposed method can start from a flat or even a random embedding cost and then iteratively update the initial costs by exploiting the adversarial attacks until satisfactory security performances are obtained. The extensive experimental results show that our method significantly outperforms the existing nonadaptive and adaptive steganography methods and achieves state-of-the-art results. Moreover, we also provide experimental results to investigate why the proposed embedding modifications seem evenly located at all audio segments despite their different content complexities, which is contrary to the content adaptive principle widely employed in modern steganography methods.

[1]  Z. Jane Wang,et al.  A Sparse Representation-Based Wavelet Domain Speech Steganography Method , 2015, IEEE/ACM Transactions on Audio, Speech, and Language Processing.

[2]  Xianfeng Zhao,et al.  Weakening the Detecting Capability of CNN-based Steganalysis , 2018, ArXiv.

[3]  Anuradha,et al.  Enhanced LSB technique for audio steganography , 2012, 2012 Third International Conference on Computing, Communication and Networking Technologies (ICCCNT'12).

[4]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[5]  Jun Zhu,et al.  Boosting Adversarial Attacks with Momentum , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[6]  Guanshuo Xu,et al.  Deep Convolutional Neural Network to Detect J-UNIWARD , 2017, IH&MMSec.

[7]  Jessica J. Fridrich,et al.  Content-Adaptive Steganography by Minimizing Statistical Detectability , 2016, IEEE Transactions on Information Forensics and Security.

[8]  Jessica J. Fridrich,et al.  Design of adaptive steganographic schemes for digital images , 2011, Electronic Imaging.

[9]  Jessica J. Fridrich,et al.  Practical methods for minimizing embedding impact in steganography , 2007, Electronic Imaging.

[10]  Jessica J. Fridrich,et al.  Universal distortion function for steganography in an arbitrary domain , 2014, EURASIP Journal on Information Security.

[11]  Kai Liu,et al.  Spatial Image Steganography Based on Generative Adversarial Network , 2018, ArXiv.

[12]  Sid-Ahmed Selouani,et al.  Speech steganography using wavelet and Fourier transforms , 2012, EURASIP J. Audio Speech Music. Process..

[13]  A. A. Zaidan,et al.  A novel embedding method to increase capacity and robustness of low-bit encoding audio steganography technique using noise gate software logic algorithm , 2010 .

[14]  Jing Dong,et al.  SSGAN: Secure Steganography Based on Generative Adversarial Networks , 2017, PCM.

[15]  Bin Li,et al.  A new cost function for spatial image steganography , 2014, 2014 IEEE International Conference on Image Processing (ICIP).

[16]  George Danezis,et al.  Generating steganographic images via adversarial training , 2017, NIPS.

[17]  Bin Li,et al.  CNN-Based Adversarial Embedding for Image Steganography , 2019, IEEE Transactions on Information Forensics and Security.

[18]  Marc Chaumont,et al.  Adaptive steganography by oracle (ASO) , 2013, 2013 IEEE International Conference on Multimedia and Expo (ICME).

[19]  Qingzhong Liu,et al.  Derivative-based audio steganalysis , 2011, TOMCCAP.

[20]  Kejiang Chen,et al.  Adversarial Examples Against Deep Neural Network based Steganalysis , 2018, IH&MMSec.

[21]  Bin Li,et al.  A Novel Image Steganography Method via Deep Convolutional Generative Adversarial Networks , 2018, IEEE Access.

[22]  Marc Chaumont,et al.  Deep Learning in steganography and steganalysis from 2015 to 2018 , 2019, Digital Media Steganography.

[23]  Diqun Yan,et al.  Audio Steganalysis with Improved Convolutional Neural Network , 2019, IH&MMSec.

[24]  Bolin Chen,et al.  Audio Steganalysis with Convolutional Neural Network , 2017, IH&MMSec.

[25]  Xianfeng Zhao,et al.  Adaptive Spatial Steganography Based on Probability-Controlled Adversarial Examples , 2018, ArXiv.

[26]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[27]  Jessica J. Fridrich,et al.  Designing steganographic distortion using directional filters , 2012, 2012 IEEE International Workshop on Information Forensics and Security (WIFS).

[28]  Rui Yang,et al.  Improved Audio Steganalytic Feature and Its Applications in Audio Forensics , 2018, ACM Trans. Multim. Comput. Commun. Appl..

[29]  Jessica J. Fridrich,et al.  Minimizing Additive Distortion in Steganography Using Syndrome-Trellis Codes , 2011, IEEE Transactions on Information Forensics and Security.

[30]  Samy Bengio,et al.  Adversarial examples in the physical world , 2016, ICLR.

[31]  Marc Chaumont,et al.  Technical points about adaptive steganography by oracle (ASO) , 2012, 2012 Proceedings of the 20th European Signal Processing Conference (EUSIPCO).

[32]  Biswanath Mukherjee,et al.  A Novel Audio Steganalysis Based on High-Order Statistics of a Distortion Measure with Hausdorff Distance , 2008, ISC.

[33]  Bin Li,et al.  Automatic Steganographic Distortion Learning Using a Generative Adversarial Network , 2017, IEEE Signal Processing Letters.

[34]  Joan Bruna,et al.  Intriguing properties of neural networks , 2013, ICLR.

[35]  Yue Zhang,et al.  Adaptive Audio Steganography Based on Advanced Audio Coding and Syndrome-Trellis Coding , 2017, IWDW.

[36]  John Klein,et al.  Exploiting Adversarial Embeddings for Better Steganography , 2019, IH&MMSec.

[37]  Rainer Böhme,et al.  Game Theory and Adaptive Steganography , 2016, IEEE Transactions on Information Forensics and Security.

[38]  Sangita Roy,et al.  Audio steganography using LSB encoding technique with increased capacity and bit error rate optimization , 2012, CCSEIT '12.

[39]  Bolin Chen,et al.  Enhancing JPEG Steganography using Iterative Adversarial Examples , 2019, 2019 IEEE International Workshop on Information Forensics and Security (WIFS).

[40]  R. Sridevi,et al.  Journal of Theoretical and Applied Information Technology Efficient Method of Audio Steganography by Modified Lsb Algorithm and Strong Encryption Key with Enhanced Security , 2022 .

[41]  Dulal C. Kar,et al.  A multi-threshold based audio steganography scheme , 2015, J. Inf. Secur. Appl..

[42]  Alan L. Yuille,et al.  Improving Transferability of Adversarial Examples With Input Diversity , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).