基于人耳听觉系统的变换域音频水印算法研究
|
摘要
数字水印技术作为一种崭新的信息安全技术,是针对数字作品的版权保护而提出来的。它的基本思想是在数字图像、音频和视频等数字产品中嵌入秘密信息,以保护数字产品的版权、证明产品的真实可靠性、跟踪盗版行为或者提供产品的附加信息。其中的秘密信息可以是版权标志、用户序列号或者是产品相关信息。
本文研究以音频信号为载体的数字水印技术,但是目前这方面的研究存在一些问题,如同步问题、能否盲检测、水印是否有意义等。文章着眼于上述某些不足,着重从以下几个方面进行研究:
首先,本文回顾了数字水印的发展历史,阐述了课题的研究背景和意义,结合人耳听觉系统的特性,介绍了数字音频水印的基本原理、传统算法及评价标准。
其次,结合心理声学模型在变换域提出了三种数字音频水印算法,水印是一幅32*32的有意义二值图像。三种算法分别简述如下:
(?)基于人耳听觉系统的FFT域音频水印算法首次对人类听觉系统(HAS)模型进行研究,算法先对音频信号进行分段离散快速傅立叶变换,再根据MPEG的心理声学模型Ⅰ计算各段的音调、噪音以及各段的总体掩蔽阈值,在低频子带中选择大于掩蔽阈值的音调分量嵌入水印。
(?)基于神经网络的近似心理声学模型音频水印算法考虑到实时性的应用,对前面算法的计算复杂度进行了降低,使用BP神经网络来估计水印嵌入强度控制因子,水印的嵌入选择在音频的每个子带的DCT域内进行。
(?)基于听觉感知模型的小波域音频水印算法综合前两种算法的优缺点,提出了一种新的基于小波包分解和心理声学模型的自适应算法,将整个音频信号的频带划分为更合理的不等宽子带,计算掩蔽阈值,由掩蔽阈值确定水印嵌入强度控制因子,水印的嵌入选择小波包分解的中频系数,并使用相关检测的方法实现了盲提取。
最后给出了三种算法的仿真实验结果,从计算量、鲁棒性、不可感知性、判决门限等方面对各种算法进行了比较和分析。实验结果表明,三种算法都具有较好的鲁棒性,不可感知性和安全性,水印嵌入容量符合版权保护的要求,是可行的数字音频水印算法。
To overcome protection against copyright problem, digital watermarking technology has been found to be a promising solution. It is a technique by which some secret information is embedded into the host signal (text, audio, image or video) that is imperceptible to human senses. The secret information may be copyright symbols, user's license or some information relative to products.
This paper study digital watermarking on audio signals. But former study exists some problems, such as synchronization, blind or non-blind detection, the watermark significant or not .This paper aims at the above defects, mainly contains the several parts as follows.
Firstly, this paper introduces the development of digital watermarking, talks about the background and significance of the study. Then it introduces the principle of digital watermarking, some traditional algorithms and evaluation standards.
Secondly, it proposes three algorithms in transfer domain based on HAS and psychoacoustic mode of MPEG. The watermark is a 32*32 significant binary image. It contains the several parts as follows.
Chapter 3 works on HAS features, propose a novel watermarking algorithm. The first step is to use FFT transform on audio signals, then find tonal components and non-tonal components and compute the global mask threshold. It chooses tonal components which are greater than mask threshold in low frequency sub-band to embed the watermark. It also brings image scrambling and pseudorandom sequence to increase the safety of the non-blind detection algorithm.
Chapter 4 considers the application of real time and decrease the compute complexity. It proposes a DCT domain watermarking scheme based on artificial neural network which can effectively estimate the watermarking scale factors for each sub-band of a frame of audio signals. Then the watermarking scale factor is used in the process of embedding. Watermarking scale factors can ensure inaudibility and increases the robustness against various attacks.
Chapter 5 analyze the advantages and disadvantages of the algorithms in the last two chapters, propose an adaptive watermarking based on wavelet Packet Transform and psychoacoustic mode which can get more reasonable sub-bands. The watermark strength is properly adjusted based on weighting factors derived from the proposed psychoacoustic model. The watermark is embedded in the middle frequency coefficients. It realizes blind detection by using relative detect solution and has good inaudibility and robustness.
Finally, comparisons and analysis among the three algorithms are given from compute complexity, inaudibility, robustness and so on. Experimental results show that the three algorithms can preserve inaudibility and robust enough to against the different signal processing operations. In addition, the watermarking information is enough to copyright protection . All the three algorithms are feasible.
本文研究以音频信号为载体的数字水印技术,但是目前这方面的研究存在一些问题,如同步问题、能否盲检测、水印是否有意义等。文章着眼于上述某些不足,着重从以下几个方面进行研究:
首先,本文回顾了数字水印的发展历史,阐述了课题的研究背景和意义,结合人耳听觉系统的特性,介绍了数字音频水印的基本原理、传统算法及评价标准。
其次,结合心理声学模型在变换域提出了三种数字音频水印算法,水印是一幅32*32的有意义二值图像。三种算法分别简述如下:
(?)基于人耳听觉系统的FFT域音频水印算法首次对人类听觉系统(HAS)模型进行研究,算法先对音频信号进行分段离散快速傅立叶变换,再根据MPEG的心理声学模型Ⅰ计算各段的音调、噪音以及各段的总体掩蔽阈值,在低频子带中选择大于掩蔽阈值的音调分量嵌入水印。
(?)基于神经网络的近似心理声学模型音频水印算法考虑到实时性的应用,对前面算法的计算复杂度进行了降低,使用BP神经网络来估计水印嵌入强度控制因子,水印的嵌入选择在音频的每个子带的DCT域内进行。
(?)基于听觉感知模型的小波域音频水印算法综合前两种算法的优缺点,提出了一种新的基于小波包分解和心理声学模型的自适应算法,将整个音频信号的频带划分为更合理的不等宽子带,计算掩蔽阈值,由掩蔽阈值确定水印嵌入强度控制因子,水印的嵌入选择小波包分解的中频系数,并使用相关检测的方法实现了盲提取。
最后给出了三种算法的仿真实验结果,从计算量、鲁棒性、不可感知性、判决门限等方面对各种算法进行了比较和分析。实验结果表明,三种算法都具有较好的鲁棒性,不可感知性和安全性,水印嵌入容量符合版权保护的要求,是可行的数字音频水印算法。
To overcome protection against copyright problem, digital watermarking technology has been found to be a promising solution. It is a technique by which some secret information is embedded into the host signal (text, audio, image or video) that is imperceptible to human senses. The secret information may be copyright symbols, user's license or some information relative to products.
This paper study digital watermarking on audio signals. But former study exists some problems, such as synchronization, blind or non-blind detection, the watermark significant or not .This paper aims at the above defects, mainly contains the several parts as follows.
Firstly, this paper introduces the development of digital watermarking, talks about the background and significance of the study. Then it introduces the principle of digital watermarking, some traditional algorithms and evaluation standards.
Secondly, it proposes three algorithms in transfer domain based on HAS and psychoacoustic mode of MPEG. The watermark is a 32*32 significant binary image. It contains the several parts as follows.
Chapter 3 works on HAS features, propose a novel watermarking algorithm. The first step is to use FFT transform on audio signals, then find tonal components and non-tonal components and compute the global mask threshold. It chooses tonal components which are greater than mask threshold in low frequency sub-band to embed the watermark. It also brings image scrambling and pseudorandom sequence to increase the safety of the non-blind detection algorithm.
Chapter 4 considers the application of real time and decrease the compute complexity. It proposes a DCT domain watermarking scheme based on artificial neural network which can effectively estimate the watermarking scale factors for each sub-band of a frame of audio signals. Then the watermarking scale factor is used in the process of embedding. Watermarking scale factors can ensure inaudibility and increases the robustness against various attacks.
Chapter 5 analyze the advantages and disadvantages of the algorithms in the last two chapters, propose an adaptive watermarking based on wavelet Packet Transform and psychoacoustic mode which can get more reasonable sub-bands. The watermark strength is properly adjusted based on weighting factors derived from the proposed psychoacoustic model. The watermark is embedded in the middle frequency coefficients. It realizes blind detection by using relative detect solution and has good inaudibility and robustness.
Finally, comparisons and analysis among the three algorithms are given from compute complexity, inaudibility, robustness and so on. Experimental results show that the three algorithms can preserve inaudibility and robust enough to against the different signal processing operations. In addition, the watermarking information is enough to copyright protection . All the three algorithms are feasible.
引文
[1]Ehsan,Tavakoli,Bijan Vosoughi Vahdat,Mohammad Bagher Shamsollahi,Reza Sameni.Audio Watermarking for Convert Communication Through Telephone System.IEEE International Symposium on Signal Processing and Information Technology,2006:955-959
[2]赵春晖,李福昌.数字音频水印技术回溯与展望.哈尔滨工程大学学报,2002,23(6):57-61
[3]张宇.音频数字水印技术发展概况及新进展.世界专业音响及灯光,2006,(2):42-47
[4]M.Arnold,M.Schmucker,and S.Wolthusen.Techniques and Applications of Digital Watermarking and Content Protection.Artech House.Boston.MA.USA,2003
[5]赵春晖,李福昌.基于小波变换的自适应音频数字水印算法[J].应用科技,2002,29(5):14-16
[6]孙圣和,陆哲明,牛夏牧等著.数字水印技术及应用.第一版。北京:科学技术出版社,2004
[7]朱丽,郭从良.心理声学模型在数字音频中的应用.电声技术,2002,8(206):11-14
[8]胡金艳,张太镒,陆从德,张春梅.一种基于神经网络的小波域音频水印算法.西安交通大学学报,2003,37(4):355-358
[9]Swanson M,Zhu B,Tewfik A,etal.Robust audio watermarking using perceptual masking[J].Signal Process,1998,66(3):337-355
[10]Ted Painter,Andree Spanias.Perceptual Coding of Digital Audio.Proceedings of the IEEE,2000,88(4):462-542
[11]Shi-Hao Wang,Yuan-Pei Lin.Wavelet Tree Quantization for Copyright Protection Watermarking.IEEE TRANSACTIONS ON IMAGE PROCESSING,2004,13(2):154-165
[12]谭建国,张文军.小波域心理听觉模型.哈尔滨工业大学学报,2007,39(11):1837-1840
[13]梁华庆,赵丽丽,钮心忻等。一种基于心理声学模型的小波域音频数字水印算法.石油大学学报,2003,27(6):112-115
[14]P Noll.MPEG Digital Audio Coding[J].IEEE SIGNAL PROCESSING MAGZINE,1997,14(9):51-89
[15]贾骏,王朔中,张新鹏.一种基于人耳听觉特性的数字音频水印方案.西安电子科技大学学报(自然科学版),2004,31(4):652-656
[16]冉启文.小波分析方法与应用.哈尔滨:哈尔滨工业大学出版社,1995
[17]何冬梅,高文.基于小波包和心理声学模型的音频编码算法.计算机研究与发展,2000,37(3):329-335
[18]李东明,王典洪,陈分雄,黄小辉.基于小波矩和神经网络检测的鲁棒水印算法.计算机应用,2006,26(8):1833-1836
[19]J.Huang,Y.Wang,Y.Q.Shi.A Blind Audio Watermarking Algorithm with Self-Synchronization.IEEE International Conference on Circuits and Systems,2002:627-630
[20]Michael Arnold.Attacks on Digital Audio Watermarks and Countermeasures.Proceedings of the Third International Conference WEB Delivering of Music (WEDELMUSIC'03),IEEE,2003:1-8
[21]I.J.Cox,M.L.Miller,and J.A.Bloom.Digital Watermarking,The Morgan Kaufmann Series in Multimedia Information and Systems.Morgan Kaufmann Publishers,San Francisco,CA,USA,2002
[22]王让定,杜呈透.一个基于HAS的音频水印检测改进算法.计算机技术与应用进展,2004:441-444
[23]Xiangyang Wang,Wei Qi,and Panpan Niu.A New Adaptive Digital Audio Watermarking Based on Support Vector Regression.IEEE TRANSACTIONS ON AUDIO,SPEECH,AND LANGUAGE PROCESSING,2007,15(8):2270-2277
[24]Changsheng Xu,Jiankang Wu,Qibin Sun.DIGITAL AUDIO WATERMARKING AND ITS APPLICATION IN MULTIMEDIA DATABASE.Fifth International Symposium on Signal Processing and its Applications,1999:91-94
[25]王剑.林福宗,基于人工神经网络的数字音频水印算法.小型微型计算机系统,2004,25(11):2006-2010
[26]丛爽.面向MATLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,1998
[27]飞思科技产品研发中心著.神经网络理论与MATLAB7实现.第一版.北京:电子工业出版社,2005
[28]Huijuan Yang,Jagdish C.Patra,Choong W.Chan.An Artificial Neural Network-based Scheme for Robust Watermarking of Audio Signals.IEEE,2002:1029-1032
[29]San Ratanasanya,Sukhanitha Poomdaeng,Suwat Tachaphetpiboon,Thumrongrat Amornraksa.New Psychoacoustic Models for Avelet Based Audio Watermarking.Proceedings of ISCIT,2005:582-585
[30]Shijun Xiang,Jiwu Huang,Senior Member,Histogram-Based Audio Watermarking Against Time-Scale Modification and Cropping Attacks.IEEE TRANSACTIONS ON MULTIMEDIA,2007,9(7):1357-1372
[31]姚惠明,杨义先,钮心忻.基于矩阵乘法共享方案的数字水印分存算法.信号处理, 2003,19(4):300-303
[32]宋坤,杨晓元,刘佳,潘峰.基于小波包统计量的音频隐秘检测算法.计算机工程与应用,2007,43(34):295-300
[33]吴平,梁青.一种基于信息分存和混沌序列的小波域信息隐藏方法.系统仿真学报,2006,18(9):2555-2558
[34]Ehsan Tavakoli,Bijan Vosoughi Vahdat,Mohammad Bagher Shamsollahi,Reza Sameni.Audio Watermarking for Convert Communication Though Telephone System.IEEE International Symposium on Signal Processing and Information Technology,2006:955-959
[35]Xiangyang Wang,Wei Qi,and Panpan Niu.A New Adaptive Digital Audio Watermarking Based on Support Vector Regression.IEEE TRANSACTIONS ON AUDIO,SPEECH,AND LANGUAGE PROCESSING,2007,15(8):2270-2277
[36]ZHANG Li,CHEN Li-min,QIAN Gong-bin.Self-synchronization Adaptive Blind Audio Watermarking,.IEEE,2006:381-384
[37]Wang Lanxun,Yin Chao,Pang Jiao.An Audio Watermark Embedding Algorithm Based on Mean-Quantization in Wavelet Domain.The Eighth International Conference on Electronic Measurement and Instruments,2007:423-425
[38]Shahrzad Esmaili,Sridhar Krishnan,Kaamran Raahemifar.Audio Watermarking using time-frequency characteristics.Can.J.Elect.Comput.Eng,2003,28(2):57-61
[39]G.Cano Rodriguez,M.Nakano Miyatake,H.M.Perez Meana.Analysis of Audio Watermarking Schemes.CIE,2005:17-20
[40]In-Kwon YEO,Hyoung Joong KIM.Modified patchwork algorithm:a novel audio watermarking scheme.IEEE Transactions on Speech and Audio processing.2003,11(4):381-386
[41]J.Huang,Y.Wang,and Y.Q.Shi.A blind audio watermarking algorithm with self-synchronization.Proc.of IEEE,Int.Sym.on Circuits,and Systems,vol.3,2002:627-630
[2]赵春晖,李福昌.数字音频水印技术回溯与展望.哈尔滨工程大学学报,2002,23(6):57-61
[3]张宇.音频数字水印技术发展概况及新进展.世界专业音响及灯光,2006,(2):42-47
[4]M.Arnold,M.Schmucker,and S.Wolthusen.Techniques and Applications of Digital Watermarking and Content Protection.Artech House.Boston.MA.USA,2003
[5]赵春晖,李福昌.基于小波变换的自适应音频数字水印算法[J].应用科技,2002,29(5):14-16
[6]孙圣和,陆哲明,牛夏牧等著.数字水印技术及应用.第一版。北京:科学技术出版社,2004
[7]朱丽,郭从良.心理声学模型在数字音频中的应用.电声技术,2002,8(206):11-14
[8]胡金艳,张太镒,陆从德,张春梅.一种基于神经网络的小波域音频水印算法.西安交通大学学报,2003,37(4):355-358
[9]Swanson M,Zhu B,Tewfik A,etal.Robust audio watermarking using perceptual masking[J].Signal Process,1998,66(3):337-355
[10]Ted Painter,Andree Spanias.Perceptual Coding of Digital Audio.Proceedings of the IEEE,2000,88(4):462-542
[11]Shi-Hao Wang,Yuan-Pei Lin.Wavelet Tree Quantization for Copyright Protection Watermarking.IEEE TRANSACTIONS ON IMAGE PROCESSING,2004,13(2):154-165
[12]谭建国,张文军.小波域心理听觉模型.哈尔滨工业大学学报,2007,39(11):1837-1840
[13]梁华庆,赵丽丽,钮心忻等。一种基于心理声学模型的小波域音频数字水印算法.石油大学学报,2003,27(6):112-115
[14]P Noll.MPEG Digital Audio Coding[J].IEEE SIGNAL PROCESSING MAGZINE,1997,14(9):51-89
[15]贾骏,王朔中,张新鹏.一种基于人耳听觉特性的数字音频水印方案.西安电子科技大学学报(自然科学版),2004,31(4):652-656
[16]冉启文.小波分析方法与应用.哈尔滨:哈尔滨工业大学出版社,1995
[17]何冬梅,高文.基于小波包和心理声学模型的音频编码算法.计算机研究与发展,2000,37(3):329-335
[18]李东明,王典洪,陈分雄,黄小辉.基于小波矩和神经网络检测的鲁棒水印算法.计算机应用,2006,26(8):1833-1836
[19]J.Huang,Y.Wang,Y.Q.Shi.A Blind Audio Watermarking Algorithm with Self-Synchronization.IEEE International Conference on Circuits and Systems,2002:627-630
[20]Michael Arnold.Attacks on Digital Audio Watermarks and Countermeasures.Proceedings of the Third International Conference WEB Delivering of Music (WEDELMUSIC'03),IEEE,2003:1-8
[21]I.J.Cox,M.L.Miller,and J.A.Bloom.Digital Watermarking,The Morgan Kaufmann Series in Multimedia Information and Systems.Morgan Kaufmann Publishers,San Francisco,CA,USA,2002
[22]王让定,杜呈透.一个基于HAS的音频水印检测改进算法.计算机技术与应用进展,2004:441-444
[23]Xiangyang Wang,Wei Qi,and Panpan Niu.A New Adaptive Digital Audio Watermarking Based on Support Vector Regression.IEEE TRANSACTIONS ON AUDIO,SPEECH,AND LANGUAGE PROCESSING,2007,15(8):2270-2277
[24]Changsheng Xu,Jiankang Wu,Qibin Sun.DIGITAL AUDIO WATERMARKING AND ITS APPLICATION IN MULTIMEDIA DATABASE.Fifth International Symposium on Signal Processing and its Applications,1999:91-94
[25]王剑.林福宗,基于人工神经网络的数字音频水印算法.小型微型计算机系统,2004,25(11):2006-2010
[26]丛爽.面向MATLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,1998
[27]飞思科技产品研发中心著.神经网络理论与MATLAB7实现.第一版.北京:电子工业出版社,2005
[28]Huijuan Yang,Jagdish C.Patra,Choong W.Chan.An Artificial Neural Network-based Scheme for Robust Watermarking of Audio Signals.IEEE,2002:1029-1032
[29]San Ratanasanya,Sukhanitha Poomdaeng,Suwat Tachaphetpiboon,Thumrongrat Amornraksa.New Psychoacoustic Models for Avelet Based Audio Watermarking.Proceedings of ISCIT,2005:582-585
[30]Shijun Xiang,Jiwu Huang,Senior Member,Histogram-Based Audio Watermarking Against Time-Scale Modification and Cropping Attacks.IEEE TRANSACTIONS ON MULTIMEDIA,2007,9(7):1357-1372
[31]姚惠明,杨义先,钮心忻.基于矩阵乘法共享方案的数字水印分存算法.信号处理, 2003,19(4):300-303
[32]宋坤,杨晓元,刘佳,潘峰.基于小波包统计量的音频隐秘检测算法.计算机工程与应用,2007,43(34):295-300
[33]吴平,梁青.一种基于信息分存和混沌序列的小波域信息隐藏方法.系统仿真学报,2006,18(9):2555-2558
[34]Ehsan Tavakoli,Bijan Vosoughi Vahdat,Mohammad Bagher Shamsollahi,Reza Sameni.Audio Watermarking for Convert Communication Though Telephone System.IEEE International Symposium on Signal Processing and Information Technology,2006:955-959
[35]Xiangyang Wang,Wei Qi,and Panpan Niu.A New Adaptive Digital Audio Watermarking Based on Support Vector Regression.IEEE TRANSACTIONS ON AUDIO,SPEECH,AND LANGUAGE PROCESSING,2007,15(8):2270-2277
[36]ZHANG Li,CHEN Li-min,QIAN Gong-bin.Self-synchronization Adaptive Blind Audio Watermarking,.IEEE,2006:381-384
[37]Wang Lanxun,Yin Chao,Pang Jiao.An Audio Watermark Embedding Algorithm Based on Mean-Quantization in Wavelet Domain.The Eighth International Conference on Electronic Measurement and Instruments,2007:423-425
[38]Shahrzad Esmaili,Sridhar Krishnan,Kaamran Raahemifar.Audio Watermarking using time-frequency characteristics.Can.J.Elect.Comput.Eng,2003,28(2):57-61
[39]G.Cano Rodriguez,M.Nakano Miyatake,H.M.Perez Meana.Analysis of Audio Watermarking Schemes.CIE,2005:17-20
[40]In-Kwon YEO,Hyoung Joong KIM.Modified patchwork algorithm:a novel audio watermarking scheme.IEEE Transactions on Speech and Audio processing.2003,11(4):381-386
[41]J.Huang,Y.Wang,and Y.Q.Shi.A blind audio watermarking algorithm with self-synchronization.Proc.of IEEE,Int.Sym.on Circuits,and Systems,vol.3,2002:627-630