基于均值量的多媒体数字音频内嵌水印算法

Multimedia Digital Audio Embedded Watermarking Algorithm Based on Mean Value

与图片或视频相比,音频是在时间域上连续变化的信号,具有更高的维度和复杂性,且需要应对各种攻击手段,为了有效提升多媒体数字音频水印的安全性,提出一种基于倒谱变换的多媒体数字音频水印算法。通过改进小波域阈值滤波算法对多媒体数字音频信号展开分解处理,经过分解处理后,对选定的音频段展开倒谱变换,获取各段音频信号对应的倒谱系数,经过计算获取倒谱系数均值,并对倒谱系数均值量化处理,最后通过修改后的倒谱系数在多媒体数字音频内嵌入水印。通过大量实验分析证明,所提方法可以承受不同类型的攻击,有效确保多媒体数字音频的安全性。

Compared with images or videos,aud io is a continuously varying signal in the time domain,characterized by higher dimensions and complexity.To effectively enhance the security of multimed ia digital audio watermarking,this article presented a multimedia digital audio watermarking algorithm based on ce pstral transformation.Firstly,an improved wavelet domain threshold filteri ng algorithm was employed to decompose the mu ltimedia digital audio signal.After the decomposition,cepstral trans form was applied to the selected audio segmen ts,thus obtaining the cepstral coefficients corresponding to each segment of the audio signal.Then,the mean value of the cepstral coefficients was calculated and quantized.Finally,the water mark was embedded in the multimedia digital a udio using the modified cepstral coefficients.Extensive experiment results valid that the proposed method can withstand different types of attacks and effectively ensure the security of multimedia digital audio.