一种心理声学模型的自同步音频水印方案  

Audio watermarking scheme with self-synchronization based on psychoacoustic model

作　　者：李秀滢[1] 段晓毅[1] 王建新[1] 

机构地区：[1]北京电子科技学院电子信息工程系,北京100070

出　　处：《计算机工程与应用》2013年第8期96-99,155,共5页Computer Engineering and Applications

基　　金：北京市自然基金项目(No.4102056);中办重点实验室青年基金(No.yqnj1004)

摘　　要：提出了一种鲁棒的数字音频水印方案,该方案主要特点:(1)应用双重置乱机制,保证水印的安全性;(2)利用心理声学模型算法,确定水印嵌入强度;(3)应用同步机制,实现水印的自同步检测;(4)在离散小波变换(DWT)域嵌入水印,提高水印的抗攻击能力;(5)利用高效的嵌入算法,提高水印的检测效率。仿真实验表明,该方案不但具有良好的不可感知性,还对诸如重采样、重量化、叠加噪声、低通滤波、MP3压缩等攻击具有较好的鲁棒性。This paper introduces a robust digital audio watermarking scheme. The features of the scheme are as below . （ 1 ）using dual-scrambling mechanisms to ensure the security of the watermark; （2） using psychoacoustic models to determine the watermark embedding strength; （3）using synchronization mechanisms to achieve watermark self-synchronize detection; （4）using DWT domain watermarking to improve anti-attack capability; （5）using efficient embedding algorithm to improve the efficiency of watermark detection. Simulating experimental results show that the scheme not only has excellent imperceptible feature, but also can stably withstand various attaches such as re-sampling, re-quantifying, noise adding, low pass filtering and MP3 compressing.

关 键 词：音频水印 心理声学模型 小波变换 同步 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]