基于DWT的可恢复数字语音取证算法

Authentication and Recovery Algorithm for Digital Speech Signal Based on DWT

提出一种可恢复的数字语音取证算法。首先给出基于DWT的数字语音压缩和重构方法,测试该方法压缩和重构语音信号的效果。然后将帧号和压缩语音作为水印进行嵌入。对被攻击的含水印信号用帧号定位被攻击的内容,压缩信号重构被攻击的内容,实施篡改恢复。与常见的语音取证算法相比,本文所提算法除对被攻击内容的篡改定位外,还能够对被攻击内容进行篡改恢复。

In this paper,an authentication and recovery algorithm was proposed for digital speech signal based on DWT.Firstly,a digital speech compression and reconstruction method was presented and the effect of this method on the compression and reconstruction of speech signal was tested.Then,the frame number and the compressed signal were embedded as watermark into the original speech.For watermark speech that was attacked,the frame number was used to locate the contents of the attack and compressed signal was used to reconstruct the attacked content and to implement tamper recovery.Compared with the common speech forensics algorithms,the scheme proposed in this paper has the ability of tamper location and tamper recovery.