基于自动选择编码及动态选词策略的文本隐写方法

doi:10.11959/j.issn.1000-436x.2022179

摘要/Abstract

摘要：

针对文本编码方式不灵活以及候选词增加导致生成的隐写文本质量较低的问题，提出了一种基于自动选择编码及动态选词策略的文本隐写方法。所提方法基于 Transformer 的神经机器翻译模型生成隐写译文。在生成隐写译文的过程中，采用定长编码和哈夫曼编码建立候选词与码字之间的映射关系，通过计算隐写词元与正常词元的概率差异百分比，实现基于概率差异阈值的动态选词。最后，比较生成的 2 种隐写译文 Sacrebleu 的大小，实现编码方式的自动选择。实验结果表明，所提方法能够生成流畅度高、可读性强的隐写译文。当隐藏容量为11.19%时，隐写译文的Sacrebleu达到10.53。

关键词: 信息隐藏, 自然语言生成, 文本隐写, 机器翻译

Abstract:

A text steganography method based on automatic selection coding and dynamic word selection strategy was proposed for the inflexible text coding method and candidate word increasing number leading to the low quality of generated steganographic text.Steganographic translations was generated based on Transformer’s neural machine translation model.In generating steganographic translations, fixed-length coding and Huffman coding were used to establish the mapping relationship between candidate words and codewords, and dynamic word selection based on the probability difference threshold was achieved by calculating the probability difference percentage between steganographic words and normal words.Finally, the size of the two generated steganographic translations Sacrebleu was compared to realize the automatic selection of coding mode.The experimental results show that the proposed method can generate steganographic translations with high fluency and readability.When the embedding rate is 11.19%, the Sacrebleu of the steganographic translation reaches 10.53.

Key words: information hiding, natural language generation, text steganography, machine translation

中图分类号:

TP309

李晖, 金家立, 金纾羽, 马卫娇. 基于自动选择编码及动态选词策略的文本隐写方法[J]. 通信学报, 2022, 43(9): 240-253.

Hui LI, Jiali JIN, Shuyu JIN, Weijiao MA. Text steganography method based on automatic selection coding and dynamic word selection strategy[J]. Journal on Communications, 2022, 43(9): 240-253.

图/表 15

图1

图2

图3

图4

图5

表1

表2

表3

表4

表5

表6

图6

图7

图8

表7

参考文献 40

[1]	LI F Y , YU Z L , QIN C ． GAN-based spatial image steganography with cross feedback mechanism［J］． Signal Processing, 2022,190:108341．
[2]	姚远志, 王锋, 严文博 ,等．基于二维码和可逆可视水印的图像隐私保护方案［J］．通信学报, 2019,40(11): 65-75．
	YAO Y Z , WANG F , YAN W B ,et al． Image privacy preservation scheme based on QR code and reversible visible watermarking［J］． Journal on Communications, 2019,40(11): 65-75．
[3]	EL-SHAHED R A , AL-BERRY M N , EBEID H M ,et al． Multi-resolution video steganography technique based on stationary wavelet transform (SWT) and singular value decomposition (SVD)［C］// International Conference on Innovative Computing and Communications． Berlin:Springer, 2022: 157-169．
[4]	PATEL R , LAD K , PATEL M ． FFT-based robust video steganography over non-dynamic region in compressed domain［M］． Berlin: Springer, 2022．
[5]	牛盼盼, 杨思宇, 王丽 ,等．基于稳健特征点的平稳小波域数字水印算法［J］．通信学报, 2019,40(11): 187-198．
	NIU P P , YANG S Y , WANG L ,et al． Digital watermarking algorithm in SWT domain based on robust local feature［J］． Journal on Communications, 2019,40(11): 187-198．
[6]	YING K Y , WANG R D , LIN Y Z ,et al． Adaptive audio steganography based on improved syndrome-trellis codes［J］． IEEE Access, 2021,9: 11705-11715．
[7]	YANG Z L , GUO X Q , CHEN Z M ,et al． RNN-stega:linguistic steganography based on recurrent neural networks［J］． IEEE Transactions on Information Forensics and Security, 2018,14(5): 1280-1295．
[8]	ZIEGLER Z M , DENG Y T , RUSH A M ． Neural linguistic steganography［J］． arXiv Preprint,arXiv:1909．01496, 2019．
[9]	ROY S , MANASMITA M ． A novel approach to format based text steganography［C］// Proceedings of the 2011 International Conference on Communication,Computing ＆ Security． New York:ACM Press, 2011: 511-516．
[10]	TANG X , CHEN M S ． Design and implementation of information hiding system based on RGB［C］// Proceedings of the 3rd International Conference on Consumer Electronics,Communications and Networks． Piscataway:IEEE Press, 2013: 217-220．
[11]	XIANG L Y , WANG X H , YANG C F ,et al． A novel linguistic steganography based on synonym run-length encoding［J］． IEICE Transactions on Information and Systems, 2017,100(2): 313-322．
[12]	MERAL H M , SANKUR B , ?ZSOY A S , ,et al． Natural language watermarking via morphosyntactic alterations［J］． Computer Speech ＆Language, 2009,23(1): 107-125．
[13]	SHNIPEROV A N , NIKITINA K A ． A text steganography method based on Markov chains［J］． Automatic Control and Computer Sciences, 2016,50(8): 802-808．
[14]	LUO Y B , HUANG Y F , LI F F ,et al． Text steganography based on ci-poetry generation using Markov chain model［J］． KSII Transactions on Internet and Information Systems (TIIS), 2016,10(9): 4568-4584．
[15]	YANG Z L , JIN S Y , HUANG Y F ,et al． Automatically generate steganographic text based on Markov model and huffman coding［J］． arXiv Preprint arXiv:1811．04720, 2018．
[16]	FANG T , JAGGI M , ARGYRAKI K ． Generating steganographic text with LSTMs［J］． arXiv Preprint,arXiv:1705．10742, 2017．
[17]	LUO Y B , HUANG Y F ． Text steganography with high embedding rate:using recurrent neural networks to generate Chinese classic poetry［C］// Proceedings of the 5th ACM Workshop On Information Hiding and Multimedia Security． New York:ACM Press, 2017: 99-104．
[18]	TONG Y J , LIU Y L , WANG J ,et al． Text steganography on RNN-generated lyrics［J］． Mathematical Biosciences and Engineering, 2019,16(5): 5451-5463．
[19]	YANG Z L , GONG B T , LI Y M ,et al． Graph-Stega:semantic controllable steganographic text generation guided by knowledge graph［J］． arXiv Preprint,arXiv:2006．08339, 2020．
[20]	YANG Z L , ZHANG P Y , JIANG M Y ,et al． RITS:real-time interactive text steganography based on automatic dialogue model［C］// International Conference on Cloud Computing and Security． Berlin:Springer, 2018: 253-264．
[21]	WAYNER P ． Mimic functions［J］． Cryptologia, 1992,16(3): 193-214．
[22]	CHAPMAN M , DAVIDA G ． Hiding the hidden:a software system for concealing ciphertext as innocuous text［C］// International Conference on Information and Communications Security． Berlin:Springer, 1997: 335-345．
[23]	BENGIO Y , DUCHARME R , VINCENT P ,et al． A neural probabilistic language model［J］． Journal of machine learning research, 2003,3: 1137-1155．
[24]	YANG Z L , HUANG Y F , ZHANG Y J ． TS-CSW:text steganalysis and hidden capacity estimation based on convolutional sliding windows［J］． Multimedia Tools and Applications, 2020,79(25): 18293-18316．
[25]	LI H , JIN S Y ． Text steganalysis based on capsule network with dynamic routing［J］． IETE Technical Review, 2021,38(1): 72-81．
[26]	GROTHOFF C , GROTHOFF K , ALKHUTOVA L ,et al． Translation-based steganography［C］// International Workshop on Information Hiding． Berlin:Springer, 2005: 219-233．
[27]	STUTSMAN R , GROTHOFF C , ATALLAH M ,et al． Lost in just the translation［C］// Proceedings of the 2006 ACM Symposium on Applied Computing． New York:ACM Press, 2006: 338-345．
[28]	MENG P , SHI Y Q , HUANG L S ,et al． LinL:lost in n-best list［C］// International Workshop on Information Hiding． Berlin:Springer, 2011: 329-341．
[29]	WANG R , UTIYAMA M , FINCH A ,et al． Sentence selection and weighting for neural machine translation domain adaptation［J］． IEEE/ACM Transactions on Audio,Speech,and Language Processing, 2018,26(10): 1727-1741．
[30]	GAIN B , HAQUE R , EKBAL A ． Not all contexts are important:the impact of effective context in conversational neural machine translation［C］// Proceedings of International Joint Conference on Neural Networks． Piscataway:IEEE Press, 2021: 1-8．
[31]	VASWANI A , SHAZEER N , PARMAR N ,et al． Attention is all you need［C］// Advances in Neural Information Processing Systems． Massachusetts:MIT Press, 2017: 5998-6008．
[32]	BRITZ D , GOLDIE A , LUONG M T ,et al． Massive exploration of neural machine translation architectures［J］． arXiv Preprint,arXiv:1703．03906, 2017．
[33]	PAPINENI K , ROUKOS S , WARD T ,et al． BLEU:a method for automatic evaluation of machine translation［C］// Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics． Stroudsburg:ACL Press, 2002: 311-318．
[34]	POST M ． A call for clarity in reporting BLEU scores［J］． arXiv Preprint,arXiv:1804．08771, 2018．
[35]	尉爽生, 杨忠良, 江旻宇 ,等．基于神经机器翻译的文本隐写方法［J］．应用科学学报, 2020,38(6): 976-985．
	YU S S , YANG Z L , JIANG M Y ,et al． Text steganography based on neural machine translation［J］． Journal of Applied Sciences, 2020,38(6): 976-985．
[36]	MURPHY B , VOGEL C ． The syntax of concealment:reliable methods for plain text information hiding［C］// Security,Steganography,and Watermarking of Multimedia Contents IX． Bellingham:SPIE Press, 2007: 351-362．
[37]	TOPKARA M , TOPKARA U , ATALLAH M J ． Information hiding through errors:a confusing approach［C］// Security,Steganography,and Watermarking of Multimedia Contents IX． Bellingham:SPIE Press, 2007: 321-332．
[38]	CHEN X Y , SUN H Y , TOBE Y ,et al． Coverless information hiding method based on the Chinese mathematical expression［C］// International Conference on Cloud Computing and Security． Berlin:Springer, 2015: 133-143．
[39]	ZHOU Z L , MU Y , ZHAO N S ,et al． Coverless information hiding method based on multi-keywords［C］// International Conference on Cloud Computing and Security． Berlin:Springer, 2016: 39-47．
[40]	WU N , YANG Z L , YANG Y ,et al． STBS-Stega:coverless text steganography based on state transition-binary sequence［J］． International Journal of Distributed Sensor Networks,2020:doi．org/10．1177/155014 7720914257．

BS	α=0	α=0.2	α=0.4	α=0.6	α=0.8	α=1.0
2	27.23	27.29	27.32	27.32	27.33	27.30
4	27.21	27.28	27.38	27.41	27.40	27.41
8	27.20	27.29	27.37	27.39	27.36	27.41
16	26.78	26.93	27.05	27.14	27.18	27.23
32	26.44	26.60	26.81	26.94	26.98	27.07

BS	bpw			Sacrebleu
BS	FLC	VLC	FVLC	FLC	VLC	FVLC
2	1.00	1.00	1.00	14.17	14.27	16.50
4	2.00	1.83	1.89	8.19	10.33	11.89
8	3.00	2.65	2.75	5.23	8.16	9.26
16	4.00	3.46	3.61	3.74	6.77	7.59
32	5.00	4.26	4.46	2.82	5.73	6.41

方法	未嵌入秘密信息	嵌入秘密信息
文献[35]方法	16	9.6
本文方法	27.60	16.72

BS	bpw			ER
BS	FLC	VLC	FVLC	FLC	VLC	FVLC
2	1.00	1.00	1.00	2.95%	2.95%	2.95%
4	2.00	1.83	1.89	5.89%	5.40%	5.56%
8	3.00	2.65	2.75	8.81%	7.83%	8.12%
16	4.00	3.46	3.61	11.74%	10.22%	10.64%
32	5.00	4.26	4.46	14.69%	12.62%	13.17%

方法	隐藏容量
文献[36]方法	0.30%
文献[27]方法	0.33%
文献[37]方法	0.35%
文献[28]方法	0.42%
文献[38]方法	1.0%
文献[35]方法	1.32%
文献[39]方法	1.57%
文献[7]方法	2.45%
文献[40]方法	2.71%
本文方法（BS=2）	2.95%
本文方法（BS=4）	5.56%