采用复倒谱和子串匹配的音频指纹算法研究

Research on audio fingerprint algorithm adopting complex cepstrum and substring matching

提出了一种基于复倒谱变换和子串匹配的数字音频指纹算法.该算法通过应用混沌方法生成用户指纹,并将原始音频信号分帧并实施复倒谱变换后,进行指纹嵌入.指纹的识别采用子串匹配的方法,计算出待检测的指纹与原始指纹之间的匹配结果,最终确定待检测的未知音频.实验结果表明:该算法具有良好的鲁棒性,能抵御常见的音频攻击,可应用于音乐版权保护、音频识别等领域.

An audio fingerprint algorithm was presented in which the user＇s fingerprinting was created by chaotic method and the initial audio signal was divided into several segments. Each segment was transformed into complex cepstrum, and the fingerprinting was embedded in the corresponding location. Substring matching method was used for fingerprinting identification, by which it is calculated if the fingerprinting to be detected matches with the original one, so that the unknown audio may be determined. Experimental results show that the algorithm is robust and able to resist many common attacks,so it can be used for copyright protection and audio identification.