|本期目录/Table of Contents|

[1]易国洪,牛小青,张胜蓝.基于可逆网络结构的自适应鲁棒图像信息隐藏算法[J].武汉工程大学学报,2025,47(05):556-564.[doi:10.19843/j.cnki.CN42-1779/TQ.202505017]
　YI Guohong,NIU Xiaoqing,ZHANG Shenglan.Adaptive robust image information hiding algorithm based on an invertible network structure[J].Journal of Wuhan Institute of Technology,2025,47(05):556-564.[doi:10.19843/j.cnki.CN42-1779/TQ.202505017]
点击复制

基于可逆网络结构的自适应鲁棒图像信息隐藏算法
(/HTML)
分享到：

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
47
期数:
2025年05期
页码:
556-564
栏目:
智能制造
出版日期:
2025-10-31

文章信息/Info

Title:
Adaptive robust image information hiding algorithm based on an invertible network structure
文章编号:
1674 - 2869（2025）05 - 0556 - 09
作者:
智能机器人湖北省重点实验室（武汉工程大学）,武汉工程大学计算机科学与工程学院,湖北 武汉 430205

Author(s):
Hubei Key Laboratory of Intelligent Robot（Wuhan Institute of Technology）；School of Computer Science & Engineering, Wuhan Institute of Technology, Wuhan 430205, China

关键词:
Keywords:
分类号:
TP301.6
DOI:
10.19843/j.cnki.CN42-1779/TQ.202505017
文献标志码:
A
摘要:
针对现有编码器-解码器结构的图像信息隐藏方法在安全性、嵌入后失真及抗噪声攻击能力上的不足，提出了一种基于可逆网络结构的自适应鲁棒图像信息隐藏算法。算法采用可逆隐藏块同时实现图像隐藏与恢复，通过集成轻量级卷积注意力模块，融合空间和通道注意力生成注意力权重，使网络根据权重自适应调整区域嵌入权重，在载体图像复杂纹理区域隐藏更多秘密信息。通过离散小波变换将信息嵌入至图像频域，通过目标损失指导在高频子带区域隐藏更多秘密信息。为进一步增强算法鲁棒性，添加了噪声层和信息增强模块，对含密图像施加高斯噪声和JPEG压缩，以模拟真实传输环境，并在恢复图像前利用信息增强模块对失真图像特征进行补偿。实验结果显示，所提出的算法在DIV2K数据集和COCO数据集上的平均峰值信噪比分别达到了44.82和41.03 dB，在应对传输环境高斯噪声和JPEG压缩情况下仍能有效恢复秘密图像，且在隐写分析残差网络（SRNet）中，检测准确率为55.01%，明显低于其他对比方法。该算法在保持视觉不可见性的同时增强了鲁棒性，为实际应用中的安全隐蔽通信提供了新的技术实现路径。
Abstract:
To address the limitations of existing encoder-decoder-based image information hiding methods in terms of security, post-embedding distortion, and resistance to noise attacks, an adaptive robust image information hiding algorithm based on an invertible network structure was proposed. The algorithm employs an invertible hidden block to simultaneously achieve image hiding and recovery. By integrating a lightweight convolutional attention module that combines spatial and channel attention, attention weights were generated, enabling the network to adaptively adjust regional embedding strength.?This allowed more secret information to be hidden in complex texture regions of the carrier image.?The information was embedded into the frequency domain of the image via discrete wavelet transform, with target loss guidance used to conceal more secret information in high-frequency subbands. To further enhance the robustness of the algorithm, a noise layer and an information enhancement module were incorporated. Gaussian noise and JPEG compression were applied to the secret-containing image to simulate real transmission conditions, while the information enhancement module compensated for distorted image features prior to recovery. Experimental results showed that the proposed algorithm achieved average peak signal-to-noise ratios of 44.82 and 41.03 dB on the DIV2K dataset and COCO dataset, respectively. It effectively recovered secret images under the Gaussian noise and JPEG compression, and attained a detection accuracy of 55.01% with the steganalysis residual network (SRNet), which was significantly lower than that of other comparative methods. The algorithm enhances robustness while maintaining visual imperceptibility, offering a new technical path for secure covert communication in practical applications.

参考文献/References:

[1] 张卫明, 王宏霞, 李斌, 等. 多媒体隐写研究进展[J]. 中国图象图形学报, 2022, 27(6): 1918-1943.
[2] CHAN C K, CHENG L M. Hiding data in images by simple LSB substitution[J]. Pattern Recognition, 2004, 37(3): 469-474.
[3] TAMIMI A A, ABDALLA A M, AL-ALLAF O. Hiding an image inside another image using variable-rate steganography[J]. International Journal of Advanced Computer Science and Applications, 2013, 4(10): 18-21.
[4] GUPTA S, GUJRAL G, AGGARWAL N. Enhanced least significant bit algorithm for image steganography[J]. International Journal of Computational Engineering & Management, 2012, 15(4): 40-42.
[5] KHEDMATI Y, PARVAZ R, BEHROO Y. 2D Hybrid chaos map for image security transform based on framelet and cellular automata[J]. Information Sciences, 2020, 512: 855-879.
[6] JAMAL S S, FARWA S, ALKHALDI A H, et al. A robust steganographic technique based on improved chaotic-range systems[J]. Chinese Journal of Physics, 2019, 61: 301-309.
[7] EMAD E, SAFEY A, REFAAT A, et al. A secure image steganography algorithm based on least significant bit and integer wavelet transform[J]. Journal of Systems Engineering and Electronics, 2018, 29(3): 639-649.
[8] ZHU J R, KAPLAN R, JOHNSON J, et al. HiDDeN: hiding data with deep networks[C]// Lecture Notes in Computer Science. Berlin: Springer, 2018: 682-697.
[9] ur REHMAN A, RAHIM R, NADEEM S. End-to-end trained CNN encoder-decoder networks for image steganography [C]// Lecture Notes in Computer Science. Berlin: Springer, 2019: 723-729.
[10] ZHANG K A, CUESTA-INFANTE A, XU L, et al. SteganoGAN: high capacity image steganography with GANs[Z/OL]. (2019-01-30) [2025-06-27]. https://arxiv.org/abs/1901.03892.
[11] JING J P, DENG X, XU M, et al. HiNet: deep image hiding by invertible network[C]// Proceedings of the IEEE International Conference on Computer Vision. Piscataway: IEEE, 2021: 4713-4722.
[12] LU S P, WANG R, ZHONG T, et al. Large-capacity image steganography based on invertible neural networks[C]// Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021:10811-10820.
[13] GUAN Z Y, JING J P, DENG X, et al. DeepMIH: deep invertible network for multiple image hiding[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(1): 372-390.
[14] XU Y M, MOU C, HU Y J, et al. Robust invertible image steganography[C]// Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 7865-7874.
[15] YANG H, XU Y T, LIU X H, et al. PRIS: practical robust invertible network for image steganography[J]. Engineering Applications of Artificial Intelligence, 2024, 133: 108419.
[16] 段新涛, 白鹿伟, 徐凯欧, 等. 基于卷积神经网络的轻量高效图像隐写[J]. 应用科学学报, 2025, 43(1): 80-93.
[17] 刘连山, 黄瑜. 基于三通道深度融合技术的图像隐写方法[J]. 信息安全研究, 2025, 11(3): 257-264.
[18] DINH L, KRUEGER D, BENGIO Y. NICE: non-linear independent components estimation[Z/OL]. (2015-04-10) [2025-06-27]. https://arxiv.org/abs/1410.8516.
[19] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]// Lecture Notes in Computer Science. Berlin: Springer, 2018: 3-19.
[20] HUANG G, LIU Z, PLEISS G, et al. Convolutional networks with dense connectivity[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(12): 8704-8716.
[21] AGUSTSSON E, TIMOFTE R. NTIRE 2017 challenge on single image super-resolution: dataset and study[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway: IEEE, 2017: 1122-1131.
[22] LIN T Y,MAIRE M, BELONGIE S, et al. Microsoft COCO: common objects in context[C]//Lecture Notes in Computer Science. Berlin: Springer, 2014: 740-755.
[23] KINGMA D P, BA J L. Adam: a method for stochastic optimization[Z/OL]. (2017?01?22) [2025?06?27]. https://arxiv.org/abs/1412.6980.
[24] 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.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期：2025-05-23
基金项目：武汉工程大学研究生教育创新基金（CX2023321）
作者简介：易国洪，硕士，教授。Email: 178035026@qq.com
引文格式：易国洪，牛小青，张胜蓝. 基于可逆网络结构的自适应鲁棒图像信息隐藏算法［J］. 武汉工程大学学报，2025，47（5）：556-564.

更新日期/Last Update: 2025-11-03