基于听觉模型的自适应音频数字水印盲检算法

提出了分段同步的音频数字水印盲检新算法 .利用伪随机序列PN(pseudo randomnoise)扩频和二相相移键控BPSK(Bi phaseshiftkeying)技术调制水印 ,并基于听觉模型生成不可听觉的水印 .水印在频域嵌入 ,提取时不需要原始音频 ,仅需与嵌入时相同的PN序列密钥即可 .提出的抵制去同步攻击的分段同步音频水印检测方法 ,可有效提取和恢复嵌入的有意义二值水印图像 .实验结果表明 ,该算法对一般信号处理类操作都具鲁棒性 ,并可有效抵制随机裁剪、MP3(MPEGaudiolayer3)压缩、替换等去同步的攻击 .

New Adaptive Audio Digital Watermarking Blind Algorithm Based on Human Auditory System

A new audio digital watermarking blind algorithm of segmentation synchronization is presented. The watermark is modulated by pseudo-random noise(PN)sequence and Bi-phase shift keying(BPSK) technique,and inaudible watermark is generated by human auditory system(HAS).The watermark is embedded in the frequency domain.When extracting the watermark,no original audio is needed,and the same PN sequence key as embedded watermark is required. The proposed segmentation synchronization scheme that resists to de-synchronized attack reduces detection error rate and improves performance.The experimental results demonstrate that the proposed watermarking algorithm is robust against some malicious de-synchronization attacks,including random cutting,MP3 compression,and replacement and so on,and the proposed algorithm can effectively recover embedded significant binary image.