基于范重心的自适应量化音频盲水印算法

Normed centre of gravity based blind adaptively quantized audio watermarking algorithm

为了实现数字音频版权保护,利用音频信号低阶统计量的稳定性和离散余弦变换的去相关性,设计了一种基于范重心的自适应量化音频盲水印算法。原始音频信号按固定长度分帧,计算分帧音频范重心坐标并确定水印嵌入候选区域,水印信息自适应量化嵌入选中的离散余弦变换中低频系数。实验结果证明:该音频水印算法具有鲁棒性和有效性,同时密钥的使用保证了算法的安全性。

In view of the low-level steady statistics of audio signal and uncorrelated property of discrete cosine transform（ DCT）,an adaptively quantized audio blind watermarking algorithm based on normed centre of gravity（NCG） is proposed here to achieve the copyright protection of digital audio.The original audio signal is framed by equalized length. The coordinates of NCGs are detected from the framed audio,and the watermark is embedded into the region which includes the NCGs. The adaptively quantized watermarking information is embedded in the low frequent coefficients of DCT domain. Experimental results show that the scheme has robustness and effectiveness and the algorithm security is guaranteed by using secret keys.