基于码书关联网络的基音调制信息隐藏检测  

Detection of Pitch Modulation Information Hiding Based on Codebook Correlation Network

作　　者：李松斌[1] 贾已真 付江云[1] 戴琼兴[1] 

机构地区：[1]中国科学院声学研究所南海研究站,海口570105 [2]海南大学信息科学技术学院,海口570228

出　　处：《计算机学报》2014年第10期2107-2117,共11页Chinese Journal of Computers

基　　金：国家自然科学基金(61303249);海南省自然科学基金(614236);海南省重大科技项目(JDJS2013006)资助

摘　　要：基音调制信息隐藏在进行基音预测时嵌入机密信息,可在低速率语音压缩编码过程中进行高隐蔽性的信息隐藏,文中试图对该种隐写进行检测.文中发现该种隐写将导致压缩语音流中相邻语音帧自适应码书的关联特性发生改变,文中以此为设计隐写分析算法的关键线索.为了量化该种关联特性,文中设计了码书关联网络模型并基于该模型得到了对隐写敏感的特征向量.最后,基于所得特征向量并结合SVM(Support Vector Machine,支持向量机)构建了隐写检测器.针对典型的低速率语音编码标准G.729以及G.723.1的实验表明,文中方法性能优于现有检测方法,实现了对基音调制信息隐藏的快速有效检测.Pitch modulation information hiding,which embeds the secret information during the pitch prediction,can hide information in low bit-rate speech codec with high imperceptibility.This paper tries to detect this type of steganography.We find that the pitch modulation steganography will change the correlation characteristics of adjacent speech frames＇ adaptive codebook in the compressed speech stream.According to this discovery,this paper adopts the codebook correlation characteristics as the key clue for the detection of the pitch modulation steganography.In order to quantify the codebook correlation characteristics,this paper designs a codebook correlation network model,and obtains the feature vector which is sensitive to the steganography.Taking the quantitative feature vector combined with supervised learning classifier,this paper builds a high performance detector.The experiments show that,for the two typical low bit-rate speech codec：G.729 and G.723.1,the proposed method has excellent performance compared to existing method and can effectively detect the pitch modulation information hiding in low bit-rate speech codec.

关 键 词：隐写分析 低速率语音编码器 基音调制信息隐藏 码书关联网络 基音预测 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]