| 205 | 0 | 81 |
| 下载次数 | 被引频次 | 阅读次数 |
针对压缩语音量化索引调制隐写分析中传统贝叶斯网络方法在低嵌入率下检测准确率不高的问题，提出了一种基于贝叶斯网络群的隐写分析方法。首先，构建用于描述语音码元自身、帧内和帧间相关性的贝叶斯网络群，通过整体样本学习构建条件概率表；其次，以每个子网络的推理结果构造个体样本的特征向量，并利用这些特征训练支持向量机(support vector machine, SVM)模型；最后，实现对未知样本的隐写分类。实验结果表明，在10 s中英文语音数据集上，以20%嵌入率进行多种隐写方法实验，所提方法的检测准确率较传统贝叶斯网络方法和深度学习方法分别提升了至少18.01个百分点和2.32个百分点。同时，检测1 s语音的平均时长为2.72 ms,满足了实时检测要求。
Abstract：To address the problem of low detection accuracy of traditional Bayesian network methods in compressed speech quantization index modulation steganalysis with low embedding rates, a steganalysis method based on Bayesian network ensembles was proposed. Firstly, Bayesian network ensembles were constructed to describe the correlations among speech codewords themselves, within frames, and between frames, and a conditional probability table was built through overall sample learning. Then, the feature vector of individual samples was constructed using the inference results of each sub-network, and these features were used to train a support vector machine(SVM) model. Finally, the steganalysis classification of unknown samples was achieved. Experimental results showed that on a 10 s Chinese and English speech dataset, with an embedding rate of 20%, this method improved the detection accuracy by at least 18.01 percentage points and 2.32 percentage points compared with traditional Bayesian network methods and deep learning methods, respectively. Moreover, the average duration for detecting 1 s of speech using this method was 2.72 ms, meeting the requirements for real-time detection.
[1] KAUR S,SINGH S,KAUR M,et al.A systematic review of computational image steganography approaches[J].Archives of computational methods in engineering,2022,29(7):4775-4797.
[2] THABIT R,UDZIR N I,YASIN S M,et al.CSNTSteg:color spacing normalization text steganography model to improve capacity and invisibility of hidden data[J].IEEE access,2022,10:65439-65458.
[3] HE S H,XU D W,YANG L,et al.Adaptive HEVC video steganography with high performance based on attention-net and PU partition modes[J].IEEE transactions on multimedia,2023,26:687-700.
[4] ELSHOUSH H T,MAHMOUD M M.Ameliorating LSB using piecewise linear chaotic map and one-time pad for superlative capacity,imperceptibility and secure audio steganography[J].IEEE access,2023,11:33354-33380.
[5] BHARTI S S,GUPTA M,AGARWAL S.A novel approach for audio steganography by processing of amplitudes and signs of secret audio separately[J].Multimedia tools and applications,2019,78(16):23179-23201.
[6] MELMAN A,EVSUTIN O.Comparative study of metaheuristic optimization algorithms for image steganography based on discrete Fourier transform domain[J].Applied soft computing,2023,132:109847.
[7] CHEN B,WORNELL G W.Quantization index modulation:a class of provably good methods for digital watermarking and information embedding[J].IEEE transactions on information theory,2001,47(4):1423-1443.
[8] XIAO B,HUANG Y F,TANG S Y.An approach to information hiding in low bit-rate speech stream[C]//IEEE GLOBECOM 2008 IEEE Global Telecommunications Conference.Piscataway:IEEE,2008:1-5.
[9] LIU P,LI S B,WANG H Q.Steganography in vector quantization process of linear predictive coding for low-bit-rate speech codec[J].Multimedia systems,2017,23(4):485-497.
[10] TIAN H,LIU J,LI S B.Improving security of quantization-index-modulation steganography in low bit-rate speech streams[J].Multimedia systems,2014,20(2):143-154.
[11] LI S B,TAO H Z,HUANG Y F.Detection of quantization index modulation steganography in G.723.1 bit stream based on quantization index sequence analysis[J].Journal of Zhejiang university SCIENCE C,2012,13(8):624-634.
[12] YANG J,LIU P,LI S B.A common method for detecting multiple steganographies in low-bit-rate compressed speech based on Bayesian inference[J].IEEE access,2019,7:128313-128324.
[13] LIN Z N,HUANG Y F,WANG J L.RNN-SM:fast steganalysis of VoIP streams using recurrent neural network[J].IEEE transactions on information forensics and security,2018,13(7):1854-1868.
[14] HU Y T,HUANG Y H,YANG Z L,et al.Detection of heterogeneous parallel steganography for low bit-rate VoIP speech streams[J].Neurocomputing,2021,419:70-79.
[15] YANG Z L,YANG H,CHANG C C,et al.Real-time steganalysis for streaming media based on multi-channel convolutional sliding windows[J].Knowledge-based systems,2022,237:107561.
[16] ZHANG C,JIANG S J,CHEN Z.TENet:leveraging transformer encoders for steganalysis of QIM steganography in VoIP speech streams[J].Multimedia tools and applications,2024,83(19):57107-57138.
[17] 赵文彬，王佳琦，吴峰，等.基于图神经网络文档相似度的实体与关系层次匹配方法[J].郑州大学学报(理学版),2023,55(6):8-14.ZHAO W B,WANG J Q,WU F,et al.A Hierarchical Matching Method of Entity and Relation Based on Graph Neural Network for Document Similarity[J].Journal of Zhengzhou university(natural science edition),2023,55(6):8-14.
[18] 彭双，伍江江，陈浩，等.基于卷积注意力网络的卫星观测任务序贯决策方法[J].郑州大学学报(理学版),2023,55(5):47-52.PENG S,WU J J,CHEN H,et al.Satellite Observation Task Sequential Decision-making Method Based on Convolutional Attention Neural Network[J].Journal of Zhengzhou university(natural science edition),2023,55(5):47-52.
[19] KHEDDAR H,HEMIS M,HIMEUR Y,et al.Deep learning for steganalysis of diverse data types:a review of methods,taxonomy,challenges and future directions[J].Neurocomputing,2024,581:127528.
[20] ANOWAR F,SADAOUI S,SELIM B.Conceptual and empirical comparison of dimensionality reduction algorithms (PCA,KPCA,LDA,MDS,SVD,LLE,ISOMAP,LE,ICA,t-SNE)[J].Computer science review,2021,40:100378.
基本信息:
DOI：10.13705/j.issn.1671-6841.2024113
中图分类号:TP391.41
引用信息:
[1]高飞鹏,杨洁.基于贝叶斯网络群的压缩语音量化索引调制隐写分析方法[J].郑州大学学报(理学版),2025,57(06):34-41.DOI:10.13705/j.issn.1671-6841.2024113.
基金信息:
浙江省自然科学基金项目(LQ20F020004)