一种抗特征分析替换攻击的数字语音取证算法

A Content Authentication Algorithm for Digital Speech Signal Robust Against Feature-Analyzed Substitution Attack

本文基于数字语音信号的保密特征,提出一种抗特征分析替换攻击的语音取证算法。给出语音信号保密特征——能量比的定义,讨论基于该特征水印嵌入的理论依据,将各帧帧号映射为整数序列的水印信息,通过量化能量比的方法嵌入到语音信号中。实验分析结果表明,本文提出的算法具有较好的不可听性和抗信号处理的能力,同时能够抵抗特征分析替换攻击,提高水印系统的安全性。

Based on the security feature of the digital speech signal, a content authentication algorithm for speech signal against feature-analyzed substitution attack was proposed.The definition of energy ratio as a se-curity feature of the speech signal was given,and the theoretical basis of the watermarking embedding is dis-cussed.Frame number was mapped to a sequence of integers and embedded by quantifying the energy ratio. The performance analysis demonstrates that the proposed schemeis inaudible,robust against feature-analyzed substitution attack,and improves the security of the watermark system.