Abstract: A novel audio information hiding algorithm based on the audio sample dots inversion is proposed.Based on extensive experiments，we find that the similarity coefficient of the original audio’s waveform and the attacked audio’s waveform is high when the frame length is appropriate.And it has little influence on the human hearing after inversing the audio sample dots in a period of time.Based on those features，at first we divide the host audio into several frames，then inverse the audio sample dots to embed secret information.We extract the secret information by judging the polarity of the similarity coefficient，and the resynchronization is designed to increase the correct rate.Experimental results show that the watermarked audio has good imperceptibility and is robust against different kinds of attacks，such as noise adding，low-pass filtering，resampling，echo，Mp3 compression，especially it can resist the attacks of the DA/AD conversions and desynchronization.

摘要： 提出了一种新颖的基于音频采样点倒置的信息隐藏算法。在一定帧长条件下，经受DA/AD转换、低通滤波等攻击以后的音频和原始音频波形相似度很高。而倒置一段时间的采样点在该段时间中不对人耳听觉产生影响。根据这些特点，算法首先对载体音频进行分帧，然后利用倒置的方法在音频中嵌入秘密信息。秘密信息通过判断相似度的极性进行提取，提取过程中算法设计了同步搜索机制，提高了秘密信息提取正确率。实验表明，嵌入秘密信息后的音频文件不仅具有良好的不可感知性，而且能够抵抗诸如低通滤波、加噪声、回声、重采样、Mp3解压缩等攻击，特别还能抵抗DA/AD转换和各种去同步攻击，具有很强的鲁棒性。