摘要
提出了一种基于能量特性分块的DCT域自适应音频水印算法。首先对音频数据做DCT变换,然后基于子块占总能量的比例将相应的频域系数分成几个部分,用每一部分的能量特性系数构造水印的嵌入强度因子,从而实现自适应的嵌入水印。实验结果表明：该算法对数字音频信号的噪声攻击、低通滤波、MPEG压缩、D/A和A/D转换等操作具有很强的鲁棒性。
A new adaptive audio watermarking scheme based on power characteristics block is proposed to offer copy protection to digital audio signals. First the audio data is converted to DCT domain, and then is separated to some blocks with power characteristics. As the power coefficient of watermark is built upon the coefficient of power characteristics block in the process of watermark embedded, the watermark can be embedded into the audio signals adaptively. Experimental results show that the proposed watermarking algorithms have high robustness to some audio signal manipulations, such as noise interfere, low pass filtering, MPEG lossy compression, D/A and A/D transform.
出处
《信号处理》
CSCD
北大核心
2006年第4期519-522,共4页
Journal of Signal Processing
基金
本课题受国家自然科学基金项目(60173044)
教育部跨世纪优秀人才培养计划基金资助
关键词
音频水印
能量特性
DCT
自适应
digital audio watermarking
power characteristic
discrete cosine transform
adaptive
作者简介
马翼平，男，1978年生，博士研究生，研究方向：音频水印，音频信息伪装。
