Text steganography on RNN-Generated lyrics

Yongju Tong; YuLing Liu; Jie Wang; Guojiang Xin; Yongju Tong; YuLing Liu; Jie Wang; Guojiang Xin

doi:10.3934/mbe.2019271

Mathematical Biosciences and Engineering

2019, Volume 16, Issue 5: 5451-5463. doi: 10.3934/mbe.2019271

Previous Article Next Article

Research article Special Issues

Text steganography on RNN-Generated lyrics

1.
College of Computer Science and Electronic Engineering, Hunan University, Changsha 410082, China
2.
Department of Computer Science, University of Massachusetts Lowell, Lowell, M.A., 01854, USA
3.
College of Management and Information Engineering, Hunan University of Chinese Medicine, Changsha 410208, China

Received: 17 January 2019 Accepted: 09 May 2019 Published: 12 June 2019

We present a Recurrent Neural Network (RNN) Encoder-Decoder model to generate Chinese pop music lyrics to hide secret information. In particular, on a given initial line of a lyric, we use the LSTM model to generate the next Chinese character or word to form a new line. In so doing, we generate the entire lyric from what has been generated so far. Using common lyric formats and rhymes we extracted, we generate lyrics embedded with secret information to meet the visual and pronunciation requirements. We carry out experiments and theoretical analysis, and show that lyrics generated by our method offer higher embedding capacities for steganography, which also look more natural than the existing steganography methods based on text generations.
- text steganography,
- lyric generation,
- recurrent neural networks,
- Char-RNN,
- Word-RNN
Citation: Yongju Tong, YuLing Liu, Jie Wang, Guojiang Xin. Text steganography on RNN-Generated lyrics[J]. Mathematical Biosciences and Engineering, 2019, 16(5): 5451-5463. doi: 10.3934/mbe.2019271

Related Papers:

Abstract

We present a Recurrent Neural Network (RNN) Encoder-Decoder model to generate Chinese pop music lyrics to hide secret information. In particular, on a given initial line of a lyric, we use the LSTM model to generate the next Chinese character or word to form a new line. In so doing, we generate the entire lyric from what has been generated so far. Using common lyric formats and rhymes we extracted, we generate lyrics embedded with secret information to meet the visual and pronunciation requirements. We carry out experiments and theoretical analysis, and show that lyrics generated by our method offer higher embedding capacities for steganography, which also look more natural than the existing steganography methods based on text generations.

References

[1]	R. H. Meng, S. G. Rice, J. Wang, et al.,A fusion steganographic algorithm based on faster R-CNN, CMC Comput. Mater. Con., 55 (2018), 1–16.
[2]	G. J. Xin, Y. L. Liu, T. Yang, et al., An adaptive audio steganography for covert wireless communication, Secur. Commun. Netw., 1 (2018), 1–10.
[3]	F. Peng, X. Q. Gong, M. Long, et al., A selective encryption scheme for protecting H.264/AVC video in multimedia social network, Multimed. Tools Appl., 76 (2018), 3235–3253.
[4]	Y. W. Kim, K. A. Moon and I. S. Oh,A text watermarking algorithm based on word classification and inter-word space statistics, International Conference on Document Analysis and Recognition, 2 (2003), 775–799.
[5]	A. M. Alattar and O. M. Alattar, Watermarking electronic text documents containing justified paragraphs and irregular line spacing, International Society for Optics and Photonics, 5306 (2004), 685–695.
[6]	B. K. Ramakrishnan, P. K. Thandra and A. V. S. M. Srinivasula, Text steganography: a novel character-level embedding algorithm using font attribute, Secur. Commun. Netw., 9 (2016), 6066– 6079.
[7]	R. Kumar, A. Malik, S. Singh, et al., A space based reversible high capacity text steganography scheme using Font type and style, International Conference on Computing, Communication and Automation, (2016), 1090–1094.
[8]	Q. Cao, X. M. Sun and L. Y. Xiang,A secure text steganography based on synonym substitution, IEEE Conference Anthology, (2014), 1–3.
[9]	L. Y. Xiang, Y. Li and W. Hao, Reversible natural language watermarking using synonym substitution and arithmetic coding, CMC Comput. Mater. Con., 55 (2018), 541–559.
[10]	J. Cong, D. Zhang and M. Pan,Chinese text information hiding based on paraphrasing technology, IEEE International Conference of Information Science and Management Engineering, 1 (2010), 39–42.
[11]	Y. Yang, Y. W. Chen and Y. L. Chen,A novel universal steganalysis algorithm based on the IQM and the SRM, CMC Comput. Mater. Con., 56 (2018), 261–271.
[12]	L. Y. Xiang, J. M. Yu, C. F. Yang, et al., A word-embedding-based steganalysis method for linguistic steganography via synonym-substitution, IEEE Access, 6 (2018), 64131–64141.
[13]	Z. S. Yu, L. S. Huang and Z. L. Chen, High embedding ratio text steganography by ci-poetry of the song dynasty, J. Chin. Inf. Proc., 23 (2009), 55–62.
[14]	J. W. Wang, T. Li, X. Y. Luo, et al., Identifying computer generated images based on quaternion central moments in color quaternion wavelet domain, IEEE. T. Circ. Syst. Vid., (2018), 1.
[15]	X. Zhang and M. Lapata,Chinese poetry generation with recurrent neural networks, International Conference on Empirical Methods in Natural Language Processing, (2014), 670–680.
[16]	Q. X. Wang, T. Y. Luo and D. Wang, Can machine generate traditional chinese poetry? A feigenbaum test, International Conference on Brain Inspired Cognitive Systems, 10023 (2016), 34–46.
[17]	Q. X. Wang, T. Y. Luo and D. Wang,Chinese song iambics generation with neural attention-based model, Association for Computing Machinery, (2016), 2943–2949.
[18]	X. Y. Yi, R. Y. Li and M. S. Sun,Generating chinese classical poems with RNN encoder-decoder, in Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data (eds. M. Sun, X. Wang, B. Chang, D. Xiong), Springer, 10565 (2017), 211–223.
[19]	Y.B. Luo, Y.F. Huang andF. F. Li,Textsteganography based onci-poetry generation usingmarkov chain model, KSII. T. Internet. Inf., 10 (2016), 4568–4584.
[20]	Y. B. Luo and Y. F. Huang, Text steganography with high embedding rate: Using recurrent neural networks to generate chinese classic poetry, 5th ACM Workshop on Information Hiding and Multimedia Security, (2017), 99–104.
[21]	C. Olah, Understanding LSTM Networks, 2015. Available from: http://colah.github.io/posts/2015-08-Understanding-LSTMs.
[22]	A. Karpathy,The unreasonable effectiveness of recurrent neural networks, 2015. Available from: http://karpathy.github.io/2015/05/21/rnn-effectiveness.
[23]	Q. Y. Du, The application of the thirteen rhymes in singing technique, Journal of Xingyi Normal University for Nationalities, (2010), in Chinese.

Reader Comments

Your name:*

Email:*
© 2019 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)