基于混沌和系数均值统计特性的音频指纹方案

Audio Fingerprint Scheme Based on Chaos and Statistical Characteristic of Average Coefficients

提出基于混沌和小波系数统计特性的数字指纹方案。通过用户信息生成混沌映射的初值,由该初值应用混沌方法生成用户指纹,并对其重复编码。对音频信号分帧并对每帧进行加汉明窗处理,对音频帧进行离散小波变换,根据小波系数统计恒定特性将指纹嵌入到每帧中。对音频帧进行逆变换,合并所有帧得到含指纹的音频信号。实验表明,该方案有较强的鲁棒性,能抵御常见音频攻击和处理方法,且对同步攻击不敏感,可用作版权保护,也可利用指纹的抗合谋攻击特性,确定叛逆者。

This paper presents a scheme of digital fingerprint in the statistical invariable characteristic of wavelet domain. Initial values of chaotic mapping are created by user information. User fingerprint is created with those values by chaotic method and coded by error-correcting coding. Audio signal is divided into frames. Use Hamming window in every frame and deal the audio frames with DWT and embed the fingerprint into each frame according to statistical invariable characteristic of wavelet coefficients. Inverse transform of audio frames are made and combined into audio signal with fingerprint. Experimental results show that the scheme is robust, it can resist common audio attacks and methods, and it is insensible to synchronization attack, so it can protect copyright. At the same time, it can apply in the confirmation of traitor because digital fingerprint has characteristics of anti-collusion.