English

应用科学学报 ›› 2025, Vol. 43 ›› Issue (6): 1015-1023.doi: 10.3969/j.issn.0255-8297.2025.06.010

• 信号与信息处理 • 上一篇    

多层级联动态嵌入的神经网络模型隐写

张恒1, 栗风永1, 秦川2   

  1. 1. 上海电力大学 计算机科学与技术学院, 上海 201306;
    2. 上海理工大学 光电信息与计算机工程学院, 上海 200093
  • 收稿日期:2025-05-25 发布日期:2025-12-19
  • 通信作者: 栗风永,教授,研究方向为人工智能安全、多媒体安全、机器学习。E-mail:fyli@shiep.edu.cn E-mail:fyli@shiep.edu.cn
  • 基金资助:
    国家自然科学基金（No. 62576203）；上海市自然科学基金（No. 20ZR1421600）

Multi-level Cascaded Dynamic Embedding Based Neural Network Model Steganography

ZHANG Heng1, LI Fengyong1, QIN Chuan2   

  1. 1. College of Computer Science and Technology, Shanghai University of Electric Power, Shanghai 201306, China;
    2. School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
  • Received:2025-05-25 Published:2025-12-19

PDF

7

摘要： 神经网络模型隐写可用于版权保护与隐蔽通信，但现有方法大多依赖模型中间层参数，导致嵌入容量有限、抗噪声能力不足。为此，本文提出一种多层级联动态嵌入的隐写方案。首先设计了多层权重分配函数，利用动态权重将秘密数据自适应地分配并嵌入到多层神经网络中，在扩展嵌入容量的同时提升了鲁棒性；然后，在模型中引入通道注意力模块，通过注意力特征强化抵消秘密数据嵌入对模型整体性能的影响。大量仿真实验表明，所提方案可以高效地解决数据隐藏的场景问题，在嵌入容量、鲁棒性、安全性方面均优于现有方法。

关键词: 模型隐写, 神经网络, 通道注意力模块, 信息隐藏

Abstract: Neural network steganography holds potential for copyright protection and covert communication. However, most existing methods rely on intermediate layer parameters of the model, which often leads to limited embedding capacity and insufficient robustness against noise. To address these issues, this paper proposes a steganographic scheme based on multi-level cascaded dynamic embedding. Firstly, a multi-layer weight distribution function is designed to adaptively allocate and embed secret data across multiple layers of the neural network using dynamic weights, thereby expanding the embedding capacity and enhancing robustness. Furthermore, a channel attention module is introduced into the model to counteract the performance degradation caused by secret data embedding, leveraging enhanced attention features to balance information embedding with model functionality preservation. Extensive simulation experiments demonstrate that the proposed scheme effectively addresses data hiding challenges and outperforms existing methods in terms of embedding capacity, robustness, and security.

Key words: model steganography, neural network, channel attention module, information hiding

中图分类号: