The Research Of Digital Audio Watermarking Theory And Key Technology Against Desynchronization Attacks

With the rapid development of network information technology,the spread of digital audio becomes more convenient and efficient.People can take advantage of digital devices so that they could deal with audio information easily,which makes the digital audio security problem becomes increasingly serious.Therefore in order to prevent digital audio from being pirated and tempered,it is very important to insert identification into digital audio.Digital watermark technology is an effective way to realize that,but the difficult point of the digital watermark technology is how to resist against desynchronization attack.Specific to this problem,two algorithms are given in the thesis.Simulation experiment results show that both of them have good imperceptibility and robustness.Main work content as follows:1.Audio watermarking algorithm based on robust QR decomposition of shearlet domain.This algorithm is proposed based on the R component invariance and the advantage of non-subsample shearlet transform(NSST).First,original audio is partitioned into audio frames,and each audio frame is segmented into two audio segments,namely audio segment A and audio segment B.Audio segment A is embed into synchronization code used to mark watermarking embedding location.Second,audio segment B mapping into two-dimensional form,then by way of NSST,one low frequency and some high frequency are achieved.In the end,with QR decomposition of low frequency sub-band and Quantization Index Modulation(QIM),the first line element of R matrix is inserted into digital watermarking.2.Audio watermarking algorithm based on robust feature point of wavelet domain.This algorithm is proposed based on content feature of audio and stability of low frequency coefficient of Undecimated Discrete Wavelet Transform(UDWT).First,original audio is dealt with by way of UDWT,then calculating first-order gradient responses of low frequency coefficient,ranking those responses in a descending order and choosing the highest response as criteria to set threshold.From this,stable and evenly distributed feature point is achieved and then setting the robust feature point as identification,inserting audio watermarking.In the end,low frequency coefficient is inserted into watermarking by way of QIM.