变换域数字水印算法的研究
摘要
随着计算机网络的飞速发展,数字媒体产品给工作和生活带来方便快捷的同时,也带来了版权安全等一系列的问题。数字水印技术作为一种有效的数字信息版权保护方法,得到了广泛的关注。
本文在分析现有数字水印技术的基础上研究了两种变换域数字水印算法:
(1)基于SVD(Singular Value Decomposition)分解的DCT(Discrete Cosine Transform)域数字水印算法。将原始图像分块并计算熵值和标准差,根据人类视觉系统特点,选取熵值较大和方差较小的块嵌入水印。对选取的原始图像块进行离散余弦变换,将二值水印图像进行混沌加密后进行奇异值分解,然后嵌入到选取块奇异值分解的S分量中。由于将水印进行了三次嵌入,加大了嵌入的水印信息量。水印提取采用了3选2的方法。仿真实验证明了算法具有良好的不可感知性,并且对有损压缩、图像添加噪声、中值滤波、剪切等图像处理操作具有较强的鲁棒性。
(2)基于图像归一化的零水印算法研究。结合图像归一化技术,研究了一种新的零水印算法。首先利用图像归一化技术在原始图像中提取构造区域,然后在重要区域中进行离散小波变换,选择低频子带进行分块离散余弦变换,提取每块直流分量组成一个矩阵并求平均值,通过判断每块系数与平均值关系得到特征信息,最后结合版权水印信息构造零水印。仿真实验表明,算法对常规处理和几何攻击均具有较好的鲁棒性。
With the rapid development of computers and networks, digital media products to work and life, while convenient, it also brings security, and other copyright issues. Digital watermarking is an effective method for copyright protection of digital information, received extensive attention:
Based on the analysis of the existing digital watermarking technique based on the proposed two digital image watermarking algorithm:
(1)Digital watermarking Algorithm Based on SVD in DCT domain. Firstly, Original image was divided into blocks and their entropy and standard deviation were calculated. According to HVS characteristics, selected some blocks with larger entropy and smaller standard deviation to embed watermark. Selected blocks are transformed into DCT domain and then operating SVD. The binary watermark image was embedded into the S component of those SVD decomposition after encrypted by chaotic encryption. Due to the watermark was embedded three times, the amount of embedded information is increased. Watermark obtained by selecting best one from three watermarks extracted from watermarked image. The experiment results show the algorithm has good invisibility and robustness to JPEG compression, image add noise, median filtering, cropping and other image processing operations.
(2)Zero-watermark Algorithm Based on Image Normalization. A new zero-watermark algorithm is proposed, combining with image normalization techniques. Firstly, the image normalization techniques are used in the original image to extract a significant-region and transform it by using DWT(Discrete Wavelet Transform), then select its low frequency sub bands to block DCT(Discrete Cosine Transform). Matrix composed by direct current component coefficients of each block extracted by determining the average relationship of low-frequency coefficients and the matrix of each block. Finally, combining with the copyright watermark to produce zero-watermark. Experimental results show that algorithm had good robustness to common image processing and resist geometric transformation.
本文在分析现有数字水印技术的基础上研究了两种变换域数字水印算法:
(1)基于SVD(Singular Value Decomposition)分解的DCT(Discrete Cosine Transform)域数字水印算法。将原始图像分块并计算熵值和标准差,根据人类视觉系统特点,选取熵值较大和方差较小的块嵌入水印。对选取的原始图像块进行离散余弦变换,将二值水印图像进行混沌加密后进行奇异值分解,然后嵌入到选取块奇异值分解的S分量中。由于将水印进行了三次嵌入,加大了嵌入的水印信息量。水印提取采用了3选2的方法。仿真实验证明了算法具有良好的不可感知性,并且对有损压缩、图像添加噪声、中值滤波、剪切等图像处理操作具有较强的鲁棒性。
(2)基于图像归一化的零水印算法研究。结合图像归一化技术,研究了一种新的零水印算法。首先利用图像归一化技术在原始图像中提取构造区域,然后在重要区域中进行离散小波变换,选择低频子带进行分块离散余弦变换,提取每块直流分量组成一个矩阵并求平均值,通过判断每块系数与平均值关系得到特征信息,最后结合版权水印信息构造零水印。仿真实验表明,算法对常规处理和几何攻击均具有较好的鲁棒性。
With the rapid development of computers and networks, digital media products to work and life, while convenient, it also brings security, and other copyright issues. Digital watermarking is an effective method for copyright protection of digital information, received extensive attention:
Based on the analysis of the existing digital watermarking technique based on the proposed two digital image watermarking algorithm:
(1)Digital watermarking Algorithm Based on SVD in DCT domain. Firstly, Original image was divided into blocks and their entropy and standard deviation were calculated. According to HVS characteristics, selected some blocks with larger entropy and smaller standard deviation to embed watermark. Selected blocks are transformed into DCT domain and then operating SVD. The binary watermark image was embedded into the S component of those SVD decomposition after encrypted by chaotic encryption. Due to the watermark was embedded three times, the amount of embedded information is increased. Watermark obtained by selecting best one from three watermarks extracted from watermarked image. The experiment results show the algorithm has good invisibility and robustness to JPEG compression, image add noise, median filtering, cropping and other image processing operations.
(2)Zero-watermark Algorithm Based on Image Normalization. A new zero-watermark algorithm is proposed, combining with image normalization techniques. Firstly, the image normalization techniques are used in the original image to extract a significant-region and transform it by using DWT(Discrete Wavelet Transform), then select its low frequency sub bands to block DCT(Discrete Cosine Transform). Matrix composed by direct current component coefficients of each block extracted by determining the average relationship of low-frequency coefficients and the matrix of each block. Finally, combining with the copyright watermark to produce zero-watermark. Experimental results show that algorithm had good robustness to common image processing and resist geometric transformation.
引文
[1] F.A.P. Petitcolas,R.J. Anderson and M.G. Kuhn. Information Hiding-A Survey[J]. Proceeding of IEEE,1999,87(7):1062-1078
[2] G. Voyatzis and I. Pitas. The Use of Watermarks in the Protection of Digital Multimedia Products[J]. Proceedings of the IEEE,1999,87(7):1197-1207
[3] R. Barnett. Digital Watermarking Applications[J]. Techniques and Challenges Electronic & Communication Engineering Journal ,1999,11(4):173-183
[4] J.M. Acken. How Watermarking Adds Value to Digital Content[C]. Communications of the ACM,1998,41(7):74-77
[5] W. Bender. Technique for Data Hiding[J]. IBM Systems Journal, 1996, 35: 313-336
[6] Ruanaith J and Pun T. Rotation scale and translation invariant digital image[J]. Proc of lCIP97, 1997: 536-539
[7] van Schyndel . R. G , Tirkel .A. Z and Osborne. C. F. A digital watermark[C].Image Proeessing , IEEE International Conference , USA ,1994:86-90.
[8]黄继武,姚若河.基于块分类的白适应图象水印算法[J].中国图象图形报,1999,4(8):640-643.
[9]温泉,孙琰锋,王树勋.基于零水印的数字水印技术研究.见:全国第三届信息隐藏学术研讨会.西安.2001.西安:西安电子科技大学出版社,2001:102～109
[10]章毓晋.图像工程(上册)图像处理(第二版)[M].清华大学出版社, 2006
[11]孙圣和,陆哲明,牛夏牧等.数字水印技术及应用[M].科学出版社, 2004
[12] M.H.Hayes.The Reconstruction of a Multidimensional Sequence from the Phase or Magnitude of the FFT[J].IEEE Transactions on Acoustics ,Speech and Signal Processing,1992(4) ,140-154
[13] I.W.Agung and P. Sweeney.Oblivious Robust Watermarking Method by Image Pre-processing[J].IEEE lectronics Letters,2000,36(14):1108-1110
[14]张小梅.基于奇异值分解和扩频技术的数字水印算法研究[D].武汉理工大学硕士学位论文,2005.
[15] Cox J, Leighton T, Shamoon T. Secure spread spectrum watermark formultimedia[J]. IEEE Transaction on Image Processing, December 1997, VOL6, No.12: 1673-1687
[16]刘九芬,黄达人,胡全军.数字水印中的正交小波基.电子与信息学报,2003,25(4):453-459
[17] D.Y.CHEN, M. Ouhyoung and J.L.Wu.A Shift-Resisting Public Watermark System for Protecting Image Processing Software[J].IEEE Transaction On Consumer Electronicds,2000,46(3):404-414
[18] Bao Paul,Xiaohu Ma.Image Adaptive Watermarking Using Wavlet Domain Singular Value Decompsition[J]. IEEE transctions on circuits and system for video technology,2005,15(1):96-102
[19]王淑琴,张金海,王卫民.一种基于奇异值分解的自适应水印算法[J].计算机仿真,2008,25(8):109-112.
[20]刘锋,孙林军.一种基于DCT和SVD的数字图像水印技术[J] .计算机应用,2005,25(8):1944- 1945.
[21]王卫卫,杨波,宋国乡。基于图像小波变换低频系数的数字水印算法。信号处理,2001,17(6):554-557
[22]曹汉强,向华.一种基于HVS的图像自适应盲水印算法[J].计算机工程与应用.2005,8:137-149
[23] S. Craver,B.-L. Yeo,M. Yeung. Technical Trial and Legal Tribulations[C].Communications of the ACM,1998,41(7):45～54
[24] J. Wood. Invariant patternre cognition: areview[J]. Pattern Recognition,1996,29(1):l-17
[25] Kunder D.Improved digital watermarking through diversity and attack characterization.Proc[C].Workshop on Multimedia Security at ACM Multimedia’99,1999:53
[26] (美) William K. Pratt著,邓鲁华,张延恒等译.数字图像处理.北京:机械工业出版社2005(1)
[27]江玉珍,杨群生.基于DCT域和DWT域的图像零水印算法的研究[J].电脑开发与应用,2005,18(6)
[28] C. Y. Lin, M. Wu, J.A. Bloom, I. J. Cox, etal. Rotation, Scale and Translation Resilient Watermarking for Images [J]. IEEE Trans. Image Process, 2001, 10(5):767–782
[29]王炳锡,陈琦,邓峰森.数字水印技术[M].西安电子科技大学出版社, 2003
[30]杜高峰.数字水印算法的研究与应用[D].硕士论文,西安建筑科技大学, 2007
[31] Barni M,et al.A DCT-domain system for robust image watermarking[J].Signal Processing,1998,66:357-372
[32]毛力,王慧琴,王燕妮等.基于分形技术与神经网络的数字水印算法[J].电视技术, 2008, 32(5): 14-16
[33]梁华庆,钮心忻,杨义先.基于数字水印的证件防伪技术.信号处理, 2004, 20(2): 143-146
[34]张旭,张贵仓.一种基于小波包分解的自适应数字水印算法[J].计算机辅助设计与图像学学报, 2007, 19(7): 931-934
[35]沈中海.基于置乱处理和几何归一化的数字水印算法研究[D].电子科技大学硕士学位论文,2008.5
[2] G. Voyatzis and I. Pitas. The Use of Watermarks in the Protection of Digital Multimedia Products[J]. Proceedings of the IEEE,1999,87(7):1197-1207
[3] R. Barnett. Digital Watermarking Applications[J]. Techniques and Challenges Electronic & Communication Engineering Journal ,1999,11(4):173-183
[4] J.M. Acken. How Watermarking Adds Value to Digital Content[C]. Communications of the ACM,1998,41(7):74-77
[5] W. Bender. Technique for Data Hiding[J]. IBM Systems Journal, 1996, 35: 313-336
[6] Ruanaith J and Pun T. Rotation scale and translation invariant digital image[J]. Proc of lCIP97, 1997: 536-539
[7] van Schyndel . R. G , Tirkel .A. Z and Osborne. C. F. A digital watermark[C].Image Proeessing , IEEE International Conference , USA ,1994:86-90.
[8]黄继武,姚若河.基于块分类的白适应图象水印算法[J].中国图象图形报,1999,4(8):640-643.
[9]温泉,孙琰锋,王树勋.基于零水印的数字水印技术研究.见:全国第三届信息隐藏学术研讨会.西安.2001.西安:西安电子科技大学出版社,2001:102～109
[10]章毓晋.图像工程(上册)图像处理(第二版)[M].清华大学出版社, 2006
[11]孙圣和,陆哲明,牛夏牧等.数字水印技术及应用[M].科学出版社, 2004
[12] M.H.Hayes.The Reconstruction of a Multidimensional Sequence from the Phase or Magnitude of the FFT[J].IEEE Transactions on Acoustics ,Speech and Signal Processing,1992(4) ,140-154
[13] I.W.Agung and P. Sweeney.Oblivious Robust Watermarking Method by Image Pre-processing[J].IEEE lectronics Letters,2000,36(14):1108-1110
[14]张小梅.基于奇异值分解和扩频技术的数字水印算法研究[D].武汉理工大学硕士学位论文,2005.
[15] Cox J, Leighton T, Shamoon T. Secure spread spectrum watermark formultimedia[J]. IEEE Transaction on Image Processing, December 1997, VOL6, No.12: 1673-1687
[16]刘九芬,黄达人,胡全军.数字水印中的正交小波基.电子与信息学报,2003,25(4):453-459
[17] D.Y.CHEN, M. Ouhyoung and J.L.Wu.A Shift-Resisting Public Watermark System for Protecting Image Processing Software[J].IEEE Transaction On Consumer Electronicds,2000,46(3):404-414
[18] Bao Paul,Xiaohu Ma.Image Adaptive Watermarking Using Wavlet Domain Singular Value Decompsition[J]. IEEE transctions on circuits and system for video technology,2005,15(1):96-102
[19]王淑琴,张金海,王卫民.一种基于奇异值分解的自适应水印算法[J].计算机仿真,2008,25(8):109-112.
[20]刘锋,孙林军.一种基于DCT和SVD的数字图像水印技术[J] .计算机应用,2005,25(8):1944- 1945.
[21]王卫卫,杨波,宋国乡。基于图像小波变换低频系数的数字水印算法。信号处理,2001,17(6):554-557
[22]曹汉强,向华.一种基于HVS的图像自适应盲水印算法[J].计算机工程与应用.2005,8:137-149
[23] S. Craver,B.-L. Yeo,M. Yeung. Technical Trial and Legal Tribulations[C].Communications of the ACM,1998,41(7):45～54
[24] J. Wood. Invariant patternre cognition: areview[J]. Pattern Recognition,1996,29(1):l-17
[25] Kunder D.Improved digital watermarking through diversity and attack characterization.Proc[C].Workshop on Multimedia Security at ACM Multimedia’99,1999:53
[26] (美) William K. Pratt著,邓鲁华,张延恒等译.数字图像处理.北京:机械工业出版社2005(1)
[27]江玉珍,杨群生.基于DCT域和DWT域的图像零水印算法的研究[J].电脑开发与应用,2005,18(6)
[28] C. Y. Lin, M. Wu, J.A. Bloom, I. J. Cox, etal. Rotation, Scale and Translation Resilient Watermarking for Images [J]. IEEE Trans. Image Process, 2001, 10(5):767–782
[29]王炳锡,陈琦,邓峰森.数字水印技术[M].西安电子科技大学出版社, 2003
[30]杜高峰.数字水印算法的研究与应用[D].硕士论文,西安建筑科技大学, 2007
[31] Barni M,et al.A DCT-domain system for robust image watermarking[J].Signal Processing,1998,66:357-372
[32]毛力,王慧琴,王燕妮等.基于分形技术与神经网络的数字水印算法[J].电视技术, 2008, 32(5): 14-16
[33]梁华庆,钮心忻,杨义先.基于数字水印的证件防伪技术.信号处理, 2004, 20(2): 143-146
[34]张旭,张贵仓.一种基于小波包分解的自适应数字水印算法[J].计算机辅助设计与图像学学报, 2007, 19(7): 931-934
[35]沈中海.基于置乱处理和几何归一化的数字水印算法研究[D].电子科技大学硕士学位论文,2008.5