【作者】 李学敏； 孔祥维； 刘航；

【Author】 Li Xuemin Kong Xiangwei Liu Hang(Information Security Research Center, Dalian University of Technology, Dalian 116024, China)

【机构】 大连理工大学信息安全研究中心；

【摘要】 为了分析音-图转换音频隐密方法嵌入的秘密信息长度,通过考察音-图转换后小波系数直方图的分布特性,提取出与嵌入容量成线性关系的峰值频率特征,建立了线性回归模型.根据音-图转换后小波系数子带选取的不同,确定不同的嵌入方案,并针对这些方案进行了相应的隐密分析.实验结果表明:所提特征与秘密信息长度之间存在显著的线性相关性;同时,对于音-图转换隐密算法性能如不可感知性、提取信息正确率等有重要影响的隐密参数TH,对回归模型参数的估计值影响较小.所提出的方法可较为准确地估计出利用音-图转换隐密方法嵌入信息的长度.更多还原

【Abstract】 In order to detect the length of secret messages embedded by audio-to-image audio steganography, via researching distribution characteristic of wavelet coefficients’ histogram after audio-to-image transform the peak frequency feature which has linear relationship with the embedding capacity is extracted, and the linear regressive model is established. According to the diverse selections of wavelet coefficients subbands after audio-to-image transform, different embedding schemes are fixed. Correspondingly,the audio steganalysis is performed for these schemes. The experimental results show that remarkable linear relativity exists between the feature proposed and secret messages length. Furthermore, the information hiding parameter TH (such as imperceptibility, accurate extraction rates, etc.) which has an important influence on the performance of the audio-to-image hiding algorithm affects the parameter estimation of regressive model less. In conclusion, the proposed algorithm can estimate the length of embedded secret messages accurately for audio-to-image audio steganography.更多还原