基于跨模态学习的鲁棒文本隐写

doi:10.3969/j.issn.0255-8297.2025.03.004

摘要/Abstract

摘要： 本文提出一种基于跨模态学习的鲁棒文本隐写方法，通过生成与图像语义一致的语句进行秘密信息的嵌入。将图像的语义特征与区域特征融合，提高文本的生成质量，并在训练阶段设计一个随机丢词的攻击层，进一步提高了隐写文本的鲁棒性。在实验部分，从抗文本攻击和抗图像攻击两个方面验证了所提出方法的鲁棒性。结果表明，所提出的隐写方法在文本生成质量与鲁棒性方面均获得了较好的性能，有效提升了隐写文本的认知隐蔽性。

关键词: 隐写, 生成式文本隐写, 跨模态, 鲁棒性

Abstract: This paper proposes a robust text steganography method based on cross-modal learning, embedding secret information by generating sentences consistent with the semantics of image. To improve text generation quality, both semantic and regional features of the image are integrated. Moreover, a random word deletion attack layer is designed during training to further enhance the robustness of the steganographic texts. Experiments evaluate the model’s robustness against both text and image attack. The results demonstrate that the proposed method achieves superior text generation quality and robustness, effectively improving the cognitive concealment of steganographic text.

Key words: steganography, generative text steganography, cross-modal, robustness

中图分类号:

TP309

马婷, 谭云, 秦姣华, 向旭宇. 基于跨模态学习的鲁棒文本隐写[J]. 应用科学学报, 2025, 43(3): 403-414.

MA Ting, TAN Yun, QIN Jiaohua, XIANG Xuyu. Robust Text Steganography Based on Cross-Modal Learning[J]. Journal of Applied Sciences, 2025, 43(3): 403-414.

参考文献

[1] Fang T, Jaggi M, Argyraki K. Generating steganographic text with LSTMs [C]//Proceedings of ACL 2017, Student Research Workshop, 2017: 100-106.
[2] 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, 2019, 14(5): 1280-1295.
[3] Yang Z L, Wei N, Liu Q H, et al. GAN-TStega: text steganography based on generative adversarial networks [C]//18th International Workshop, Digital Forensics and Watermarking (IWDW), 2019: 18-31.
[4] 秦川，王萌，司广文，等. 基于绝句生成的构造式信息隐藏算法[J]. 计算机学报, 2021, 44(4): 773-785. Qin C, Wang M, Si G W, et al. Constructive information hiding with chinese quatrain generation [J]. Chinese Journal of Computers, 2021, 44(4): 773-785. (in Chinese)
[5] Wang C L, Liu Y L, Tong Y J, et al. GAN-GLS: generative lyric steganography based on generative adversarial networks [J]. Computers, Materials & Continua, 2021, 69(1): 375-1390.
[6] 杨婉霞，陈帅，管磊，等. 宋词自动生成的信息隐藏算法[J]. 信息网络安全, 2022, 22(7): 46-54. Yang W X, Chen S, Guan L, et al. Information steganography algorithm for automatic generation of song ci [J]. Netinfo Security, 2022, 22(7): 46-54. (in Chinese)
[7] 秦川，李蓉受，钱振兴，等. 基于宋词生成的大容量构造式信息隐藏算法[J]. 计算机学报, 2023, 46(1): 17-30. Qin C, Li R S, Qian Z X, et al. Large-capacity constructive information hiding based on song ci generation [J]. Chinese Journal of Computers, 2023, 46(1): 17-30. (in Chinese)
[8] Yang Z L, Zhang S Y, Hu Y T, et al. VAE-Stega: linguistic steganography based on variational auto-encoder [J]. IEEE Transactions on Information Forensics and Security, 2020, 16: 880-895.
[9] Zhang S Y, Yang Z L, Yang J S, et al. Provably secure generative linguistic steganography [C]//Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021. Stroudsburg: Association for Computational Linguistics, 2021: 3046-3055.
[10] Zhou X J, Peng W L, Yang B Y, et al. Linguistic steganography based on adaptive probability distribution [J]. IEEE Transactions on Dependable and Secure Computing, 2022, 19(5): 2982-2997.
[11] Yi B, Wu H Z, Feng G R, et al. ALiSa: acrostic linguistic steganography based on BERT and Gibbs sampling [J]. IEEE Signal Processing Letters, 2022, 29: 687-691.
[12] Zheng X Y, Fang Y R, Wu H Z. General framework for reversible data hiding in texts based on masked language modeling [C]//2022 IEEE 24th International Workshop on Multimedia Signal Processing (MMSP), 2022: 1-6.
[13] Lu T H, Liu G S, Zhang R, et al. Neural linguistic steganography with controllable security [C]//2023 International Joint Conference on Neural Networks (IJCNN), Gold Coast, Australia, 2023: 1-8.
[14] Lin K, Luo Y Y, Zhang Z J, et al. Zero-shot generative linguistic steganography [C]//Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2024: 5168-5182.
[15] 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, 2018: 253-264.
[16] Zhang S Y, Yang Z L, Yang J S, et al. Linguistic steganography: from symbolic space to semantic space [J]. IEEE Signal Processing Letters, 2021, 28: 11-15.
[17] Yang Z L, Xiang L Y, Zhang S Y, et al. Linguistic generative steganography with enhanced cognitive-imperceptibility [J]. IEEE Signal Processing Letters, 2021, 28: 409-413.
[18] Sun B W, Li Y M, Zhang J, et al. Topic controlled steganography via graph-to-text generation [J]. CMES-Computer Modeling in Engineering and Sciences, 2023, 136(1): 157-176.
[19] 向凌云，杨双辉，王蓉，等. 基于序列到隐写序列的约束型自然语言信息隐藏方法[J]. 计算机学报, 2023, 46(8): 1650-1669. Xiang L Y, Yang S H, Wang R, et al. Constrained linguistic steganography based on sequence to steganographic sequence model [J]. Chinese Journal of Computers, 2023, 46(8): 1650-1669. (in Chinese)
[20] Yang T Y, Wu H Z, Yi B, et al. Semantic-preserving linguistic steganography by pivot translation and semantic-aware bins coding [J]. IEEE Transactions on Dependable and Secure Computing, 2024, 21(1): 139-152.
[21] Zhang R F, Liu J Y, Zhang R. Controllable semantic linguistic steganography via summarization generation [C]//2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024: 4560-4564.
[22] Wang R, Xiang L Y, Liu Y F, et al. PNG-Stega: progressive non-autoregressive generative linguistic steganography [J]. IEEE Signal Processing Letters, 2023, 30: 528-532.
[23] 刘玉玲，王翠林，付章杰. 语义空间下基于情感表达的生成式文本隐写方法[J]. 通信学报, 2023, 44(4): 176-186. Liu Y L, Wang C L, Fu Z J. Generative text steganography method based on emotional expression in semantic space [J]. Journal on Communications, 2023, 44(4): 176-186. (in Chinese)
[24] Xiang L Y, Ou C F, Zeng D J. Linguistic steganography: hiding information in syntax space [J]. IEEE Signal Processing Letters, 2024, 31: 261-265.
[25] Ding C H, Fu Z J, Yang Z L, et al. Context-aware linguistic steganography model based on neural machine translation [J]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2024, 32: 868-878.
[26] Wen J, Zhou X J, Li M D, et al. A novel natural language steganographic framework based on image description neural network [J]. Journal of Visual Communication and Image Representation, 2019, 61: 157-169.
[27] Li M D, Mu K, Zhong P, et al. Generating steganographic image description by dynamic synonym substitution [J]. Signal Process, 2019, 164: 193-201.
[28] Hu Y T, Yang Z L, Cao H, et al. Multi-modal steganography based on semantic relevancy [C]//19th International Workshop, Digital Forensics and Watermarking (IWDW), 2020: 3-14.
[29] Hu Y T, Li H Y, Song J N, et al. MM-Stega: multi-modal steganography based on text-image matching [C]//International Conference on Artificial Intelligence and Security, 2020: 313-325.
[30] Hu Y T, Cao H, Yang Z L, et al. Improving text-image matching with adversarial learning and circle loss for multi-modal steganography [C]//19th International Workshop, Digital Forensics and Watermarking (IWDW), 2020: 41-52.
[31] Guo Y Y, Wu H Z, Zhang X P. Steganographic visual story with mutual-perceived joint attention [J]. EURASIP Journal on Image and Video Processing, 2021(1): 5-19.
[32] Wang X L, Wang Y F, Chen K J, et al. ICStega: image captioning-based semantically controllable linguistic steganography [C]//2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023: 1-5.
[33] Peng W L, Wang T, Qian Z X, et al. Cross-modal text steganography against synonym substitution-based text attack [J]. IEEE Signal Processing Letters, 2023, 30: 299-303.
[34] Lin T Y, Maire M, Belongie S, et al. Microsoft COCO: common objects in context [C]//European Conference on Computer Vision, 2014: 740-755.
[35] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016: 770-778.
[36] Ren S, He K, Girshick R, et al. Faster R-CNN: towards real-time object detection with region proposal networks [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[37] Huang L, Wang W, Chen J, et al. Attention on attention for image captioning [C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019: 4633-4642.
[38] Hochreiter S, Schmidhuber J. Long short-term memory [J]. Neural Computation, 1997, 9(8): 1735-1780.