一种基于条件生成对抗网络的图像隐写方法研究与实现

doi:10.3969/j.issn.1671-1122.2021.11.006

信息网络安全 ›› 2021, Vol. 21 ›› Issue (11): 48-57.doi: 10.3969/j.issn.1671-1122.2021.11.006

一种基于条件生成对抗网络的图像隐写方法研究与实现

雷雨(), 刘佳, 李军, 柯彦

武警工程大学密码工程学院,西安 710086

收稿日期:2020-09-28 出版日期:2021-11-10 发布日期:2021-11-24
通讯作者: 雷雨 E-mail:ly1a2b3c@163.com
作者简介:雷雨（1987—）,男,湖北,讲师,硕士,主要研究方向为图像隐写与分析|刘佳（1982—）,男,河南,副教授,博士,主要研究方向为信息隐藏|李军（1987—）,男,湖南,讲师,硕士,主要研究方向为图像隐写与分析|柯彦（1991—）,男,河南,博士研究生,主要研究方向为可逆信息隐藏
基金资助:
国家自然科学基金(61872384)

Research and Implementation of a Image Steganography Method Based on Conditional Generative Adversarial Networks

LEI Yu(), LIU Jia, LI Jun, KE Yan

College of Cryptographic Engineering, Engineering University of PAP, Xi’an 710086, China

Received:2020-09-28 Online:2021-11-10 Published:2021-11-24
Contact: LEI Yu E-mail:ly1a2b3c@163.com

摘要/Abstract

摘要：

通过纹理图像合成实现隐写是一类常见的载体合成隐写方法,但纹理图像不具有语义特征,这类方法容易在多次传输后引起攻击者注意。生成对抗网络（Generative Adversarial Networks,GAN）利用博弈策略让生成器和判别器对抗,理论上训练达到最优的生成器能使生成样本的分布与真实数据相同。因此,在理想情况下用GAN实现合成隐写能构造出自然的含密图像。目前基于GAN的载体合成图像隐写方法的问题之一是无法控制生成图像的内容。针对该问题,文章提出了一种基于条件生成对抗网络的图像隐写方法,该方法用随机噪声和条件信息的组合作为隐空间的表示来训练生成器,使生成图像受条件信息控制;用生成图像和条件信息的组合作为概率空间的表示来训练提取器,使提取噪声与驱动噪声一致。实验结果表明,该方法可完成含密图像生成与消息提取的功能,在生成含密图像时能利用条件信息控制图像内容的生成,同时保证生成的含密图像质量和消息提取正确率不降低。

关键词: 生成对抗网络, 条件生成对抗网络, 图像隐写

Abstract:

Steganography by texture image synthesis was a kind of common carrier synthesis steganography method. However, the texture image did not have semantic features, so this kind of method was easy to attract the attention of attackers after multiple transmission. The generative adversarial networks used game strategy to make the generator confront the discriminator. In theory, the generator with the best training could make the distribution of the generated samples the same as the real data. In ideal conditions, using GAN to realize synthesis steganography could construct a natural image. One of the problems of the image synthesis steganography based on GAN was that it couldn’t control the content of the generated image. To solve this problem, this paper proposed an image steganography method based on conditional generative adversarial networks. In this method, the combination of random noise and condition information was used as the representation of hidden space to train the generator, so that the generated image was controlled by condition information. The combination of generated image and condition information was used as the representation of probability space to train the extractor, so that the extracted noise was consistent with the driving noise. The experimental results showed that this method could complete the function of generating the image and extracting the message. The outstanding feature was that it could use the condition information to control the content of the image. At the same time, the image quality and accuracy rate of the message extraction were close to the method in the comparison.

Key words: generative adversarial networks, conditional generative adversarial networks, image steganography

中图分类号:

TP309

雷雨, 刘佳, 李军, 柯彦. 一种基于条件生成对抗网络的图像隐写方法研究与实现[J]. 信息网络安全, 2021, 21(11): 48-57.

LEI Yu, LIU Jia, LI Jun, KE Yan. Research and Implementation of a Image Steganography Method Based on Conditional Generative Adversarial Networks[J]. Netinfo Security, 2021, 21(11): 48-57.

图/表 15

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

图12

图13

表1

表2

参考文献 20

[1]	FILLER T, JUDAS J, FRIDRICH J. Minimizing Additive Aistortion in Steganography Using Syndrome-trellis Codes[J]. IEEE Transactions on Information Forensics and Security, 2011, 6(3):920-935. doi: 10.1109/TIFS.2011.2134094 URL
[2]	SALLEE P. Model-based Methods for Steganography and Steganalysis[J]. International Journal of Image and Graphics, 2005, 5(1):167-189. doi: 10.1142/S0219467805001719 URL
[3]	PEVNY T, FILLER T, BAS P. Using High-dimensional Image Models to Perform Highly Undetectable Steganography[J]. Lecture Notes in Computer Science, 2010, 3(12):161-177.
[4]	HOLUB V, FRIDRICH J. Digital Image Steganography Using Universal Distortion[C]// ACM. The First ACM Workshop on Information Hiding and Multimedia Security, June 17-19, 2013, Montpellier, France. New York: ACM, 2013: 59-68.
[5]	HOLUB V, FRIDRICH J, DENEMARK T. Universal Distortion Function for Steganography in An Arbitrary Domain[J]. Eurasip Journal on Information Security, 2014, 1(1):1-13.
[6]	GOODFELLOW I J, POUGET A J, MIRZA M, et al. Generative Adversarial Nets[EB/OL]. https://arxiv.org/pdf/1406.2661.pdf, 2018-06-10.
[7]	TANG Weixuan, TAN Shunquan, LI Bin, et al. Automatic Stegano-graphic Distortion Learning Using A Generative Adversarial Net-Work[J]. IEEE Signal Processing Letters, 2017, 24(10):1547-1551. doi: 10.1109/LSP.2017.2745572 URL
[8]	XU Guanshuo, WU Hanzhou, SHI Yunqing. Structural Design of Convolutional Neural Networks for Steganalysis[J]. IEEE Signal Processing Letters, 2016, 23(5):708-712. doi: 10.1109/LSP.2016.2548421 URL
[9]	ZHOU Zhili, CAO Yi, SUN Xingming. Coverless Information Hiding Based on Bag-of-words Model of Image[J]. Journal of Applied Sciences, 2016, 34(5):527-536.
	周志立, 曹燚, 孙星明. 基于图像bag-of-words 模型的无载体信息隐藏[J]. 应用科学学报, 2016, 34(5):527-536.
[10]	WU Kuochen, WANG Chungming. Steganography Using Reversible Texture Synjournal[J]. IEEE Transactions on Image Processing, 2015, 24(1):130-139. doi: 10.1109/TIP.2014.2371246 pmid: 25415988
[11]	HU Donghui, WANG Liang, JIANG Wenjie, et al. A Novel Image Steganography Method via Deep Convolutional Generative[J]. IEEE Access, 2018, 99(6):38303-38314.
[12]	ZHANG Minqing, LI Zonghan, LIU Jia, et al. Generative Steganography Based on Boundary Equilibrium Generative Adversarial Network[EB/OL]. https://doi.org/10.13705/j.issn.1671-6841.2019419, 2020-01-06.
	张敏情, 李宗翰, 刘佳, 等. 基于边界平衡生成对抗网络的生成式隐写[EB/OL]. https://doi.org/10.13705/j.issn.1671-6841.2019419, 2020-01-06.
[13]	ZHANG Zhuo, LIU Jia, KE Yan, et al. Generative Steganography by Sampling[J]. IEEE Access, 2019, 96(7):118586-118597.
[14]	MIRZA M, OSINDERO S. Conditional Generative Adversarial Nets[EB/OL]. https://arxiv.org/abs/1411.1784.pdf, 2018-03-12.
[15]	LIU Jia, KE Yan, LEI Yu, et al. Application of Generative Adversarial Network in Image Steganography[J], Journal of Wuhan University(Natural Science Edition). 2019, 65(2):139-152.
	刘佳, 柯彦, 雷雨, 等. 生成对抗网络在图像隐写中的应用[J], 武汉大学学报(理学版), 2019, 65(2):139-152.
[16]	VOLKHONSKIY D, NAZAROV I, BORISENKO B, et al. Steganographic Generative Adversarial Networks[EB/OL]. http://arxiv.org/abs/1703.05502.pdf, 2018-02-12.
[17]	SHI Haichao, DONG Jing, WANG Wei, et al. SSGAN: Secure Steganography Based on Generative Adversarial Networks[EB/OL]. https://arxiv.org/abs/1707.01613v3.pdf, 2018-04-12.
[18]	WANG Yaojie, NIU Ke, YANG Xiaoyuan. Image Steganography Scheme Based on GANs[J]. Netinfo Security, 2019, 19(5):54-60.
	王耀杰, 钮可, 杨晓元. 基于生成对抗网络的图像隐藏方案[J]. 信息网络安全, 2019, 19(5):54-60.
[19]	LIU Mingming, ZHANG Minqing, LIU Jia, et al. Coverless Information Hiding Based on Generative Adversarial Networks[J]. Journal of Applied Sciences, 2018, 36(2):371-382.
	刘明明, 张敏情, 刘佳, 等. 基于生成对抗网络的无载体信息隐藏[J]. 应用科学学报, 2018, 36(2):371-382.
[20]	CHEN Lu, MAO Weiyun, SU Lei, et al. Design of Steganography Based on Generative Adversarial Networks[J]. High Technology Letters, 2019, 29(7):632-639.
	陈璐, 毛玮韵, 苏磊, 等. 基于生成对抗网络的隐写术设计[J]. 高技术通讯, 2019, 29(7):632-639.

training steps	5000	10000	15000	20000	25000	30000	35000	40000	45000	50000
文献[11]方法正确率/%	56.28	58.14	60.23	61.31	61.76	62.48	62.99	62.81	63.57	63.47
本文方法正确率/%	60.44	61.02	62.02	62.07	61.79	62.46	62.57	62.45	62.92	63.21

嵌入长度	\|s\|=100	\|s\|=120	\|s\|=150	\|s\|=200
文献[11]方法正确率/%	63.47	62.6	61.45	61.1
本文方法正确率/%	63.21	61.64	60.73	60.71

一种基于条件生成对抗网络的图像隐写方法研究与实现

Research and Implementation of a Image Steganography Method Based on Conditional Generative Adversarial Networks

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 15

参考文献 20

相关文章 3

编辑推荐

Metrics

本文评价

[1]	王耀杰, 杨晓元, 刘文超. 基于数字化卡登格的生成图像隐写方案[J]. 信息网络安全, 2021, 21(2): 70-77.
[2]	彭中联, 万巍, 荆涛, 魏金侠. 基于改进CGANs的入侵检测方法研究[J]. 信息网络安全, 2020, 20(5): 47-56.
[3]	毕新亮, 杨海滨, 杨晓元, 黄思远. 基于StarGAN的生成式图像隐写方案[J]. 信息网络安全, 2020, 20(12): 64-71.