Image steganography method based on conditional generative adversarial networks and hybrid attention mechanism

doi:10.11772/j.issn.1001-9081.2025020204

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (2): 475-484.DOI: 10.11772/j.issn.1001-9081.2025020204

• Cyber security • Previous Articles

Image steganography method based on conditional generative adversarial networks and hybrid attention mechanism

Ming LI¹^,², Mengqi WANG¹(), Aili ZHANG¹^,², Hua REN¹^,², Yuqiang DOU¹^,²

^1.School of Computer and Information Engineering，Henan Normal University，Xinxiang Henan 453007，China
^2.Henan Key Laboratory of Educational Artificial Intelligence and Personalized Learning （Henan Normal University），Xinxiang Henan 453007，China

Received:2025-03-06 Revised:2025-04-07 Accepted:2025-04-18 Online:2025-04-24 Published:2026-02-10
Contact: Mengqi WANG
About author:LI Ming， born in 1981， Ph. D.， associate professor. His research interests include information hiding， image encryption， adversarial examples.
WANG Mengqi， born in 1999， M. S. candidate. His research interests include image steganography， adversarial examples. Email:2208183007@stu.htu.edu.cn
ZHANG Aili， born in 1966， Ph. D.， professor. Her research interests include signal processing.
REN Hua， born in 1992， Ph. D.， lecturer. Her research interests include encrypted domain information hiding.
DOU Yuqiang， born in 1980， Ph. D.， lecturer. His research interests include image information security.
Supported by:
Henan Province Science and Technology Research and Development Program(252102210163);Key Scientific Research Project of Henan Higher Education Institutions(25A520028);2023 Henan Province Undergraduate Uiniversities Research-Oriented Teaching Program(Jiaogao ［2023］ 388-21)

基于条件生成对抗网络和混合注意力机制的图像隐写方法

李名¹^,², 王孟齐¹(), 张爱丽¹^,², 任花¹^,², 窦育强¹^,²

^1.河南师范大学计算机与信息工程学院，河南新乡 453007
^2.河南省教育人工智能与个性化学习重点实验室（河南师范大学），河南新乡 453007

通讯作者: 王孟齐
作者简介:李名（1981—），男，河南新乡人，副教授，博士，主要研究方向：信息隐藏、图像加密、对抗样本
王孟齐（1999—），男，河南焦作人，硕士研究生，主要研究方向：图像隐写、对抗样本 Email:2208183007@stu.htu.edu.cn
张爱丽（1966—），女，河南滑县人，教授，博士，主要研究方向：信号处理
任花（1992—），女，河南息县人，讲师，博士，主要研究方向：密文域信息隐藏
窦育强（1980—），男，河南卫辉人，讲师，博士，主要研究方向：图像信息安全。
基金资助:
2023年河南省本科高校研究性教学项目(教高［2023］388号21);河南省科技攻关计划项目(252102210163);河南省科技攻关计划项目(252102210181);河南省高等学校重点科研项目(25A520028)

Abstract

Abstract:

Current deep image steganography methods based on image-in-image concealment face challenges in practical applications of privacy protection and secure communication due to insufficient security of stego images and distortion in recovered secret images. To address these issues， a Conditional Generative Adversarial Network and Convolutional Block Attention Module-based image-in-image steganography method （CBAM-CGAN） was proposed. Firstly， a hybrid attention module was introduced into the generator network to enable the generator’s comprehensive learning of image features from both channel and spatial dimensions， thereby enhancing the visual quality of stego images. Secondly， residual connections were employed to reduce feature loss of secret images during network learning， and through adversarial training between the extractor and the discriminator， noise-free extraction of secret images was achieved. Finally， adversarial training between the generator and the steganalyzer was implemented to improve the stego image security. Experimental results on public datasets including COCO demonstrate that compared with steganography method StegGAN， the proposed steganography method achieves the Peak Signal-to-Noise Ratio （PSNR） improvements of 4.37 dB and 4.71 dB for stego and decrypted images， respectively， along with Structure Similarity Index Measure （SSIM） enhancements of 9.16% and 6.46%， respectively. For security， the proposed method has the detection Accuracy （Acc） against steganalyzer Ye-Net decreased by 9.35 percentage points with the False Negative Rate （FNR） increased by 12.01 percentage points. It can be seen that the proposed method ensures stego image security while achieving high-quality secret image recovery.

Key words: deep learning, image steganography, conditional Adversarial Generation Network (GAN), hybrid attention mechanism, image-in-image concealment

摘要：

目前以图藏图的深度隐写术存在隐写图像安全性不强以及恢复的秘密图像中存在图像失真的问题，难以实际应用于隐私保护和秘密通信。针对以上问题，提出一种基于条件生成对抗网络和混合注意力机制的以图藏图隐写方法（CBAM-CGAN）。首先，在生成器网络中引入混合注意模块，帮助生成器从通道和空间维度全面地学习图像特征，提高隐写图像的视觉质量；其次，引入残差连接降低网络学习过程中秘密图像的特征损失，并通过提取器和判别器的对抗训练，实现秘密图像的无噪声提取；最后，通过生成器和隐写分析器的对抗训练，提高隐写图像的安全性。在COCO等公开数据集上的实验结果显示，与StegGAN隐写方法相比，所提隐写方法的隐写图像和解密图像的峰值信噪比（PSNR）分别提高了4.37 dB和4.71 dB，结构相似性（SSIM）分别提高了9.16%和6.46%。在安全性方面，所提方法面对隐写分析器Ye-Net的检测，检测准确率（Acc）降低了9.35个百分点，误检率（FNR）提升了12.01个百分点。可见，所提方法在保证隐写图像安全性的同时能高质量地恢复秘密图像。

关键词: 深度学习, 图像隐写, 条件对抗生成网络, 混合注意力机制, 以图藏图

CLC Number:

TN911.73

Ming LI, Mengqi WANG, Aili ZHANG, Hua REN, Yuqiang DOU. Image steganography method based on conditional generative adversarial networks and hybrid attention mechanism[J]. Journal of Computer Applications, 2026, 46(2): 475-484.

李名, 王孟齐, 张爱丽, 任花, 窦育强. 基于条件生成对抗网络和混合注意力机制的图像隐写方法[J]. 《计算机应用》唯一官方网站, 2026, 46(2): 475-484.

Figures/Tables 12

References 39

[1]	付章杰，王帆，孙星明，等. 基于深度学习的图像隐写方法研究［J］. 计算机学报， 2020， 43（9）： 1656-1672.
	FU Z J， WANG F， SUN X M， et al. Research on steganography of digital images based on deep learning［J］. Chinese Journal of Computers， 2020， 43（9）： 1656-1672.
[2]	MIELIKAINEN J. LSB matching revisited［J］. IEEE Signal Processing Letters， 2006， 13（5）： 285-287.
[3]	WESTFELD A， PFITZMANN A. Attacks on steganographic systems： breaking the steganographic utilities EzStego， Jsteg， Steganos， and S-Tools-and some lessons learned［C］// Proceedings of the 3rd International Workshop on Information Hiding， LNCS 1768. Berlin： Springer， 2000： 61-76.
[4]	FRIDRICH J， GOLJAN M， DU R. Reliable detection of LSB steganography in color and grayscale images［C］// Proceedings of the 2001 Workshop on Multimedia and Security： New Challenges. New York： ACM， 2001： 27-30.
[5]	WU D C， TSAI W H. A steganographic method for images by pixel value differencing［J］. Pattern Recognition Letters， 2003， 24（9/10）： 1613-1626.
[6]	SWAIN G. Very high capacity image steganography technique using quotient value differencing and LSB substitution［J］. Arabian Journal for Science and Engineering， 2019， 44（4）： 2995-3004.
[7]	PEVNÝ T， FILLER T， BAS P. Using high-dimensional image models to perform highly undetectable steganography［C］// Proceedings of the 2010 International Conference on Information Hiding， LNCS 6387. Berlin： Springer， 2010： 161-177.
[8]	HOLUB V， FRIDRICH J. Designing steganographic distortion using directional filters［C］// Proceedings of the 2012 IEEE International Workshop on Information Forensics and Security. Piscataway： IEEE， 2012： 234-239.
[9]	LI B， TAN S， WANG M， et al. Investigation on cost assignment in spatial image steganography［J］. IEEE Transactions on Information Forensics and Security， 2014， 9（8）： 1264-1277.
[10]	王耀杰，钮可，杨晓元. 基于生成对抗网络的信息隐藏方案［J］. 计算机应用， 2018， 38（10）： 2923-2928.
	WANG Y J， NIU K， YANG X Y. Information hiding scheme based on generative adversarial network［J］. Journal of Computer Applications， 2018， 38（10）： 2923-2928.
[11]	ZHU J， KAPLAN R， JOHNSON J， et al. HiDDeN： hiding data with deep networks［C］// Proceedings of the 2018 European Conference on Computer Vision， LNCS 11219. Cham： Springer， 2018： 682-697.
[12]	AHMADI M， NOROUZI A， KARIMI N， et al. ReDMark： framework for residual diffusion watermarking based on deep networks［J］. Expert Systems with Applications， 2020， 146： No.113157.
[13]	ZHANG K A， CUESTA-INFANTE A， XU L， et al. SteganoGAN： high capacity image steganography with GANs［EB/OL］. ［2024-02-20］..
[14]	LUO X， ZHAN R， CHANG H， et al. Distortion agnostic deep watermarking［C］// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2020： 13545-13554.
[15]	BALUJA S. Hiding images in plain sight： deep steganography［C］// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook： Curran Associates Inc.， 2017： 2066-2076.
[16]	WU P， YANG Y， LI X. StegNet： mega image steganography capacity with deep convolutional network［J］. Future Internet， 2018， 10（6）： No.54.
[17]	DUAN X， JIA K， LI B， et al. Reversible image steganography scheme based on a U-Net structure［J］. IEEE Access， 2019， 7： 9314-9323.
[18]	ZHANG R， DONG S， LIU J. Invisible steganography via generative adversarial networks［J］. Multimedia Tools and Applications， 2019， 78（7）： 8559-8575.
[19]	MUÑOZ-RAMÍREZ D O， PONOMARYOV V， REYES-REYES R， et al. Steganographic framework for hiding a color image into digital images［C］// Proceedings of the 2019 IEEE International Scientific-Practical Conference Problems of Infocommunications， Science and Technology. Piscataway： IEEE， 2019： 63-66.
[20]	JENYNOF A， AHMAD T. Image to image steganography using U-Net architecture with MobileNet convolutional neural network［C］// Proceedings of the 14th International Conference on Computing Communication and Networking Technologies. Piscataway： IEEE， 2023： 1-7.
[21]	LIU X， MA Z， CHEN Z， et al. Hiding multiple images into a single image via joint compressive autoencoders［J］. Pattern Recognition， 2022， 131： No.108842.
[22]	ZHANG L， LU Y， LI J， et al. Deep adaptive hiding network for image hiding using attentive frequency extraction and gradual depth extraction［J］. Neural Computing and Applications， 2023， 35（15）： 10909-10927.
[23]	SINGH B， SHARMA P K， HUDDEDAR S A， et al. StegGAN： hiding image within image using conditional generative adversarial networks［J］. Multimedia Tools and Applications， 2022， 81（28）： 40511-40533.
[24]	XU G， WU H Z， SHI Y Q. Structural design of convolutional neural networks for steganalysis［J］. IEEE Signal Processing Letters， 2016， 23（5）： 708-712.
[25]	YU C. Attention based data hiding with generative adversarial networks［C］// Proceedings of the 34th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2020： 1120-1128.
[26]	ZHU X， CHENG D， ZHANG Z， et al. An empirical study of spatial attention mechanisms in deep networks［C］// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2019： 6687-6696.
[27]	JIE H， LI S， GANG S. Squeeze-and-excitation networks［C］// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2018： 7132-7141.
[28]	TAN J， LIAO X， LIU J， et al. Channel attention image steganography with generative adversarial networks［J］. IEEE Transactions on Network Science and Engineering， 2022， 9（2）： 888-903.
[29]	WOO S， PARK J， LEE J Y， et al. CBAM： convolutional block attention module［C］// Proceedings of the 2018 European Conference on Computer Vision， LNCS 11211. Cham： Springer， 2018： 3-19.
[30]	YE J， NI J， YI Y. Deep learning hierarchical representations for image steganalysis［J］. IEEE Transactions on Information Forensics and Security， 2017， 12（11）： 2545-2557.
[31]	BOROUMAND M， CHEN M， FRIDRICH J. Deep residual network for steganalysis of digital images［J］. IEEE Transactions on Information Forensics and Security， 2019， 14（5）： 1181-1193.
[32]	JOHNSON J， ALAHI A， LI F F. Perceptual losses for real-time style transfer and super-resolution［C］// Proceedings of the 2016 European Conference on Computer Vision， LNCS 9906. Cham： Springer， 2016： 694-711.
[33]	KINGMA D P， BA J L. Adam： a method for stochastic optimization［EB/OL］. ［2024-01-30］..
[34]	WANG Z， BOUIK A C， SHEIKH H R， et al. Image quality assessment： from error visibility to structural similarity［J］. IEEE Transactions on Image Processing， 2004， 13（4）： 600-612.
[35]	WANG Z， SIMONCELLI E P， BOVIK A C. Multiscale structural similarity for image quality assessment［C］// Proceedings of the 37th Asilomar Conference on Signals， Systems and Computers. Piscataway： IEEE， 2003： 1398-1402.
[36]	SHEIKH H R， BOVIK A C. Image information and visual quality［J］. IEEE Transactions on Image Processing， 2006， 15（2）： 430-444.
[37]	WANG Z， BOVIK A C. A universal image quality index［J］. IEEE Signal Processing Letters， 2002， 9（3）： 81-84.
[38]	张逸为，张卫明，俞能海. 针对特定测试样本的隐写分析方法［J］. 软件学报， 2018， 29（4）： 987-1001.
	ZHANG Y W， ZHANG W M， YU N H. Specific testing sample steganalysis［J］. Journal of Software， 2018， 29（4）： 987-1001.
[39]	ZHANG R， ZHU F， LIU J， et al. Depth-wise separable convolutions and multi-level pooling for an efficient spatial CNN-based steganalysis［J］. IEEE Transactions on Information Forensics and Security， 2020， 15： 1138-1150.

方法	图像对	SSIM/%	PSNR/dB	MS-SSIM	VIF	UQI
NIPS-17^［15］	C， S	0.848 1	31.509 0	0.951 9	0.564 5	0.952 3
NIPS-17^［15］	M， M´	0.758 1	31.485 5	0.683 6	0.785 7	0.921 1
DAH-Net^［22］	C， S	0.800 3	32.279 9	0.895 3	0.803 6	0.965 0
DAH-Net^［22］	M， M´	0.788 3	30.559 4	0.798 6	0.752 6	0.929 6
StegGAN^［23］	C， S	0.910 1	36.210 9	0.978 0	0.835 9	0.983 1
StegGAN^［23］	M， M´	0.902 8	31.923 5	0.954 2	0.661 2	0.937 6
CBAM-CGAN	C， S	0.993 5	40.582 6	0.995 6	0.872 5	0.995 6
CBAM-CGAN	M， M´	0.961 2	36.632 7	0.979 6	0.776 4	0.975 6

方法	图像对	SSIM/%	PSNR/dB	MS-SSIM	VIF	UQI
NIPS-17^［15］	C， S	0.848 1	31.509 0	0.951 9	0.564 5	0.952 3
NIPS-17^［15］	M， M´	0.758 1	31.485 5	0.683 6	0.785 7	0.921 1
DAH-Net^［22］	C， S	0.800 3	32.279 9	0.895 3	0.803 6	0.965 0
DAH-Net^［22］	M， M´	0.788 3	30.559 4	0.798 6	0.752 6	0.929 6
StegGAN^［23］	C， S	0.910 1	36.210 9	0.978 0	0.835 9	0.983 1
StegGAN^［23］	M， M´	0.902 8	31.923 5	0.954 2	0.661 2	0.937 6
CBAM-CGAN	C， S	0.993 5	40.582 6	0.995 6	0.872 5	0.995 6
CBAM-CGAN	M， M´	0.961 2	36.632 7	0.979 6	0.776 4	0.975 6

隐写方法	封面图像和隐写图像		秘密图像和提取图像
隐写方法	PSNR/dB	SSIM	PSNR/dB	SSIM
CBAM-CGAN	40.58	0.993 5	36.63	0.961 2
对照组1	34.67	0.920 6	36.28	0.941 6
对照组2	33.88	0.917 9	36.68	0.940 5
对照组3	39.59	0.976 9	31.68	0.901 7

隐写方法	封面图像和隐写图像		秘密图像和提取图像
隐写方法	PSNR/dB	SSIM	PSNR/dB	SSIM
CBAM-CGAN	40.58	0.993 5	36.63	0.961 2
对照组1	34.67	0.920 6	36.28	0.941 6
对照组2	33.88	0.917 9	36.68	0.940 5
对照组3	39.59	0.976 9	31.68	0.901 7

隐写方法	Ye-Net^［30］		SR-Net^［31］		Zhu-Net^［39］
隐写方法	Acc	FNR/%	Acc	FNR/%	Acc	FNR/%
CBAM-CGAN	0.518 9	42.52	0.512 1	40.61	0.520 6	41.57
文献［20］方法	0.581 6	32.58	0.591 5	27.89	0.567 1	29.05
StegGAN^［23］	0.612 4	30.51	0.620 1	28.45	0.581 6	30.15
对照组1	0.692 3	10.25	0.694 2	12.74	0.719 4	13.06
对照组2	0.640 2	12.51	0.571 8	11.52	0.543 1	11.46

Image steganography method based on conditional generative adversarial networks and hybrid attention mechanism

基于条件生成对抗网络和混合注意力机制的图像隐写方法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 12

References 39

Related Articles 15

Recommended Articles

Metrics

攻击类型	所提方法		文献［20］方法		文献［23］方法
攻击类型	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
平均	37.1	0.935 5	29.8	0.829 5	30.0	0.796 1
Salt and pepper	38.5	0.961 2	31.8	0.875 7	30.9	0.850 8
Poisson	36.7	0.946 5	34.1	0.853 4	32.1	0.875 7
Speckle	37.2	0.958 3	30.5	0.822 8	30.8	0.835 2
Average filter	35.9	0.924 1	28.3	0.876 7	28.4	0.828 1
Cropping（25%）	34.1	0.919 5	26.7	0.793 3	26.2	0.716 5
Cropping（50%）	32.6	0.901 8	24.8	0.777 2	30.1	0.694 8
Resize（50%）	36.4	0.923 2	28.9	0.804 6	32.5	0.779 4
Gaussian noise	37.8	0.937 6	30.1	0.821 8	31.7	0.828 5
JPEG Compression（QF=50）	35.3	0.947 3	33.4	0.840 3	27.5	0.755 7

[1]	Zeyi GUO, Fenglian LI, Lichun XU. Double decision mechanism-based deep symbolic regression algorithm [J]. Journal of Computer Applications, 2026, 46(2): 406-415.
[2]	Jinjiao LIN, Canshun ZHANG, Shuya CHEN, Tianxin WANG, Jian LIAN, Yonghui XU. Vehicle insurance fraud detection method based on improved graph attention network [J]. Journal of Computer Applications, 2026, 46(2): 437-444.
[3]	Haoqian JIANG, Dong ZHANG, Guanyu LI, Heng CHEN. SetaCRS： Conversational recommender system with structure-enhanced hierarchical task-oriented prompting strategy [J]. Journal of Computer Applications, 2026, 46(2): 368-377.
[4]	Xiaoyong BIAN, Peiyang YUAN, Qiren HU. Dual-coding space-frequency mixing method for infrared small target detection [J]. Journal of Computer Applications, 2026, 46(1): 252-259.
[5]	Zhihui ZAN, Yajing WANG, Ke LI, Zhixiang YANG, Guangyu YANG. Multi-feature fusion speech emotion recognition method based on SAA-CNN-BiLSTM network [J]. Journal of Computer Applications, 2026, 46(1): 69-76.
[6]	Panfeng JING, Yudong LIANG, Chaowei LI, Junru GUO, Jinyu GUO. Semi-supervised image dehazing algorithm based on teacher-student learning [J]. Journal of Computer Applications, 2025, 45(9): 2975-2983.
[7]	Weigang LI, Jiale SHAO, Zhiqiang TIAN. Point cloud classification and segmentation network based on dual attention mechanism and multi-scale fusion [J]. Journal of Computer Applications, 2025, 45(9): 3003-3010.
[8]	Zhixiong XU, Bo LI, Xiaoyong BIAN, Qiren HU. Adversarial sample embedded attention U-Net for 3D medical image segmentation [J]. Journal of Computer Applications, 2025, 45(9): 3011-3016.
[9]	Hongjun ZHANG, Gaojun PAN, Hao YE, Yubin LU, Yiheng MIAO. Multi-source heterogeneous data analysis method combining deep learning and tensor decomposition [J]. Journal of Computer Applications, 2025, 45(9): 2838-2847.
[10]	Jin LI, Liqun LIU. SAR and visible image fusion based on residual Swin Transformer [J]. Journal of Computer Applications, 2025, 45(9): 2949-2956.
[11]	Bing YIN, Zhenhua LING, Yin LIN, Changfeng XI, Ying LIU. Emotion recognition method compatible with missing modal reasoning [J]. Journal of Computer Applications, 2025, 45(9): 2764-2772.
[12]	Yanhua LIAO, Yuanxia YAN, Wenlin PAN. Multi-target detection algorithm for traffic intersection images based on YOLOv9 [J]. Journal of Computer Applications, 2025, 45(8): 2555-2565.
[13]	Lina GE, Mingyu WANG, Lei TIAN. Review of research on efficiency of federated learning [J]. Journal of Computer Applications, 2025, 45(8): 2387-2398.
[14]	Peng PENG, Ziting CAI, Wenling LIU, Caihua CHEN, Wei ZENG, Baolai HUANG. Speech emotion recognition method based on hybrid Siamese network with CNN and bidirectional GRU [J]. Journal of Computer Applications, 2025, 45(8): 2515-2521.
[15]	Shuo ZHANG, Guokai SUN, Yuan ZHUANG, Xiaoyu FENG, Jingzhi WANG. Dynamic detection method of eclipse attacks for blockchain node analysis [J]. Journal of Computer Applications, 2025, 45(8): 2428-2436.

隐写方法	图像对	DIV2K		ImageNet		SIPI
隐写方法	图像对	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB
NIPS-17^［15］	C， S	0.803 6	28.541 8	0.832 4	30.025 1	0.816 3	29.059 1
NIPS-17^［15］	M， M'	0.709 7	29.346 1	0.725 1	28.612 5	0.740 2	27.034 1
DAH-Net^［22］	C， S	0.793 1	31.025 8	0.775 1	30.204 7	0.759 1	29.047 2
DAH-Net^［22］	M， M'	0.753 1	28.215 2	0.736 2	27.810 4	0.739 1	29.452 1
StegGAN^［23］	C， S	0.872 8	34.128 7	0.882 1	32.621 7	0.851 4	33.512 7
StegGAN^［23］	M， M'	0.815 6	30.013 5	0.873 3	29.865 4	0.835 4	30.594 2
CBAM-CGAN	C， S	0.953 1	36.218 5	0.943 6	37.021 9	0.946 1	36.015 4
CBAM-CGAN	M， M'	0.923 6	34.598 7	0.932 1	35.237 1	0.924 6	35.651 3

隐写方法	图像对	DIV2K		ImageNet		SIPI
隐写方法	图像对	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB
NIPS-17^［15］	C， S	0.803 6	28.541 8	0.832 4	30.025 1	0.816 3	29.059 1
NIPS-17^［15］	M， M'	0.709 7	29.346 1	0.725 1	28.612 5	0.740 2	27.034 1
DAH-Net^［22］	C， S	0.793 1	31.025 8	0.775 1	30.204 7	0.759 1	29.047 2
DAH-Net^［22］	M， M'	0.753 1	28.215 2	0.736 2	27.810 4	0.739 1	29.452 1
StegGAN^［23］	C， S	0.872 8	34.128 7	0.882 1	32.621 7	0.851 4	33.512 7
StegGAN^［23］	M， M'	0.815 6	30.013 5	0.873 3	29.865 4	0.835 4	30.594 2
CBAM-CGAN	C， S	0.953 1	36.218 5	0.943 6	37.021 9	0.946 1	36.015 4
CBAM-CGAN	M， M'	0.923 6	34.598 7	0.932 1	35.237 1	0.924 6	35.651 3