摘要
倒谱分析广泛运用于音频信号处理的很多方面,但目前研究倒谱域音频水印算法的文献很少,而且现有的水印算法大多没有实现盲检测,不利于实际运用。该文提出了一种具有较高鲁棒性的混沌音频水印算法,并通过自适应神经网络模糊推理系统(AdaptiveNeuro-FuzzyInferenceSystem,ANFIS)实现了盲检测。实验证明,该算法的水印嵌入易实现,水印检测准确性高,时间代价小,具有较强的抗攻击性和较好的不可感知性,很有实用价值。
Cepstrum analysis is widely used in many aspects of audio signal processing,but present papers about audio watermarking in cepstrum domain are very few.And most existing watermark algorithms are not blind,which are not beneficial to practical applications.This paper gives a chaotic audio watermarking algorithm with good robustness.Based on ANFIS,the algorithm achieves blind detection.Experiments have shown that the algorithm of embedding is easily done,the performance in watermark detection is high,the time expense is little,the resisting to attacks is strong and the imperceptibility is good.So the algorithm is very fit for utility.
出处
《计算机工程与应用》
CSCD
北大核心
2005年第10期67-70,共4页
Computer Engineering and Applications
基金
国家部委重点科研项目
陕西省自然科学基金资助项目(编号:2001X28)
关键词
数字音频水印
混沌
ANFIS
鲁棒性
盲检测
digital audio watermarking,chaotic,ANFIS,robustness,blind detection
