基于内容的自适应量化数字音频盲水印算法

现有量化调制音频水印技术普遍采纳了均匀量化策略，其不仅需要结合大量实验确定量化步长。而且难以保证不可感知性与鲁棒性的良好平衡．以模糊聚类分析理论为基础．提出了一种自适应量化小波域数字音频盲水印算法．该算法能够结合数字音频局部特征．利用模糊聚类分析自适应确定量化步长，并在小波域内将水印信号嵌入到音频数据段的低频分量中．仿真实验表明，所提出的算法不仅具有较好的不可感知性．而且对于叠加噪声、重新采样、重新量化、低通滤波、MP3压缩等音频处理与攻击均具有较好的鲁棒性．

A Novel Blind Audio Watermarking Algorithm by Adaptive Quantization

Uniform quantization strategy has been widely adopted by currently available quantization modulation techniques of audio watermarking.It not only needs to ascertain the quantization step with massive experiments,moreover it's difficulty in guaranteeing the fine balance between the transparency and robustness.On the basis of fuzzy clustering theory,a new blind digital audio watermarking algorithm based on Discrete Wavelet Transform(DWT) is presented in this paper.The watermark can be embedded into the low frequency components of DWT domain by adaptive quantization according to the local audio correlation.The simulated experiment shows that the proposed watermarking algorithm is imperceptible and robust against various signal processing such as noise adding,re-sampling,re-quantization,low pass filtering and MPEG Audio Layer 3(MP3) compression.