数字图像水印新方法研究
|
摘要
随着计算机技术的飞速发展,数字多媒体产品市场在繁荣的同时也带来版权纠纷等严重问题。研究信息版权的保护技术具有非常重要的意义。数字水印是集信号处理、图像处理、模式识别等多种学科为一体的交叉学科,已成为版权保护使用最为广泛的技术,其中,鲁棒水印是一个非常重要的研究方向。
本文在研究分析了现有数字水印技术基础上,将各种新概念新理论与数字水印技术结合,为其带来更强的鲁棒性和不可见性。主要内容如下:
(1)提出了一种基于奇异值分解的抗几何攻击图像水印算法,算法利用了奇异值良好的稳定特性,通过对宿主载体和水印图像的奇异值进行混合运算嵌入水印,同时引入了FastICA算法,在满足算法强鲁棒性和透明性的同时,保证了快速的收敛性和盲提取要求,属于新观点创新。实验结果表明,算法能很好地抵抗各种常规攻击,对几何攻击具有优于传统小波算法无法比拟的强鲁棒性。
(2)提出了一种基于Contourlet变换抗几何攻击数字图像水印方法,主要解决现有同类方法不能抵抗几何攻击问题。在Contourlet变换域实现水印的嵌入与提取,使用几何矩方法完成水印同步,并利用Contourlet变换对遭攻击后水印进行复原。本方法具有抵抗几何攻击能力强,水印提取效果好,水印嵌入位置可自主定位的优点,适用于对抵抗几何攻击要求强的场合。
(3)基于新型Mean Shift算法快速的收敛性和稳定性,提出了一种基于Mean Shift纹理特征聚类的Contourlet域水印算法。算法使用基于灰度共生矩阵的三种纹理特征量:能量、熵、对比度,通过Mean Shift快速聚类算法提取图像强纹理区域和判断嵌入强度,并在Contourlet域变换大系数中添加水印,直接、准确、高效,自适应于各种图像;具有强的隐蔽性、抗噪声攻击能力及鲁棒性,算法实现了盲检测。
With the rapid development of computer science, the digital product market has got a highly boom, and a great number of information security problems as well. It’s very important to doing the technical research about copyright protection. Digital watermark is a kind of technique combined with signal processing, image processing, mode recognition etc. It has been one of the most popular techniques for copyright protection.
Based on the analysis of existing digital watermark techniques, this paper introduces many new theories to digital watermark algothrims. The experiments have shown that those new theories really improved the robustness and ivisiblity of the watermark. The main contents are described as followed:
(1) A blind watermarking algorithm based on the singular values decomposition for its favorable stability was proposed. This algorithm embeds a small disturbance watermarking in the image’s singular values to achieve strong robustness and transparency, and introduces the FastICA algorithm to ensure fast astringency and blind extraction. The experiment shows that this algorithm has the strong robustness to general attacks. It can also resist the geometrical attack. It shows the unique superiority and great advantage.
(2) A geometry attack resisted algorithm was proposed in contourlet domain. The watermark was embedded in the coefficients of contourlet transform. Before watermark extraction, the geometric attacked image can be reverted to its original shape by geometric moments which were caculated just after the watermark was embedded. The experiment results showed that this algorithm can resist all kinds of geometry attack, and also can resist many kinds of general attacks with a good quality of the extracted watermark. The embedding position can be located adaptively. This alogrithm is quite fit for the situation with highly requirement of resistence about geometry attack.
(3) In order to improve the robustness and efficiency of the watermark algorithm,a robust watermark algorithm in contourlet domain based on Mean Shift texture clustering is proposed in this paper for the first time. This algorithm selects the energy, entropy, contrast for texture features, rapidly and accurately extracts strong regional texture of host images by the way of Mean Shift fast clustering algorithm, which make algorithm’s the convergence rate being better than FCM clustering as well as K-means clustering. This algorithm embedded watermark into the big coefficients in contourlet domain of image strong texture area, through texture clustering outcome adaptively controls the watermark embedding locations and strength, which directly brings the more covert and strong robustness. The experiment shows that this algorithm can be very good to meet the requirements of rapid clustering, and adaptively selects the precise watermark embedding position and embedding strength. This algorithm has the strong robustness to regular attacks. This algorithm has realized the blind detection.
本文在研究分析了现有数字水印技术基础上,将各种新概念新理论与数字水印技术结合,为其带来更强的鲁棒性和不可见性。主要内容如下:
(1)提出了一种基于奇异值分解的抗几何攻击图像水印算法,算法利用了奇异值良好的稳定特性,通过对宿主载体和水印图像的奇异值进行混合运算嵌入水印,同时引入了FastICA算法,在满足算法强鲁棒性和透明性的同时,保证了快速的收敛性和盲提取要求,属于新观点创新。实验结果表明,算法能很好地抵抗各种常规攻击,对几何攻击具有优于传统小波算法无法比拟的强鲁棒性。
(2)提出了一种基于Contourlet变换抗几何攻击数字图像水印方法,主要解决现有同类方法不能抵抗几何攻击问题。在Contourlet变换域实现水印的嵌入与提取,使用几何矩方法完成水印同步,并利用Contourlet变换对遭攻击后水印进行复原。本方法具有抵抗几何攻击能力强,水印提取效果好,水印嵌入位置可自主定位的优点,适用于对抵抗几何攻击要求强的场合。
(3)基于新型Mean Shift算法快速的收敛性和稳定性,提出了一种基于Mean Shift纹理特征聚类的Contourlet域水印算法。算法使用基于灰度共生矩阵的三种纹理特征量:能量、熵、对比度,通过Mean Shift快速聚类算法提取图像强纹理区域和判断嵌入强度,并在Contourlet域变换大系数中添加水印,直接、准确、高效,自适应于各种图像;具有强的隐蔽性、抗噪声攻击能力及鲁棒性,算法实现了盲检测。
With the rapid development of computer science, the digital product market has got a highly boom, and a great number of information security problems as well. It’s very important to doing the technical research about copyright protection. Digital watermark is a kind of technique combined with signal processing, image processing, mode recognition etc. It has been one of the most popular techniques for copyright protection.
Based on the analysis of existing digital watermark techniques, this paper introduces many new theories to digital watermark algothrims. The experiments have shown that those new theories really improved the robustness and ivisiblity of the watermark. The main contents are described as followed:
(1) A blind watermarking algorithm based on the singular values decomposition for its favorable stability was proposed. This algorithm embeds a small disturbance watermarking in the image’s singular values to achieve strong robustness and transparency, and introduces the FastICA algorithm to ensure fast astringency and blind extraction. The experiment shows that this algorithm has the strong robustness to general attacks. It can also resist the geometrical attack. It shows the unique superiority and great advantage.
(2) A geometry attack resisted algorithm was proposed in contourlet domain. The watermark was embedded in the coefficients of contourlet transform. Before watermark extraction, the geometric attacked image can be reverted to its original shape by geometric moments which were caculated just after the watermark was embedded. The experiment results showed that this algorithm can resist all kinds of geometry attack, and also can resist many kinds of general attacks with a good quality of the extracted watermark. The embedding position can be located adaptively. This alogrithm is quite fit for the situation with highly requirement of resistence about geometry attack.
(3) In order to improve the robustness and efficiency of the watermark algorithm,a robust watermark algorithm in contourlet domain based on Mean Shift texture clustering is proposed in this paper for the first time. This algorithm selects the energy, entropy, contrast for texture features, rapidly and accurately extracts strong regional texture of host images by the way of Mean Shift fast clustering algorithm, which make algorithm’s the convergence rate being better than FCM clustering as well as K-means clustering. This algorithm embedded watermark into the big coefficients in contourlet domain of image strong texture area, through texture clustering outcome adaptively controls the watermark embedding locations and strength, which directly brings the more covert and strong robustness. The experiment shows that this algorithm can be very good to meet the requirements of rapid clustering, and adaptively selects the precise watermark embedding position and embedding strength. This algorithm has the strong robustness to regular attacks. This algorithm has realized the blind detection.
引文
[1]卢开澄,计算机密码学—计算机网络中的数据保密与安全,清华大学出版社,2003
[2]范科峰,莫玮,曹山,赵新华,裴庆祺,数字版权管理技术及应用研究进展,电子学报,2007,35(6):1139-1147
[3]和红杰,陈帆,自嵌入水印算法的安全性分析,电子学报,2007,35(3):557-562
[4]王炳锡,陈琦,邓峰森,数字水印技术,西安:西安电子科技大学出版社, 2003
[5]孙圣和,陆哲明,牛夏牧,数字水印技术及应用,北京:科学出版社, 2004
[6]钮心忻,信息隐藏与数字水印,北京:北京邮电大学出版社,2004
[7] Potdar. V. M., Han. S, Chang. E., A survey of digital image watermarking techniques, 2005 3rd IEEE International Conference on Industrial Informatics, 2005. INDIN '05. 10-12 Aug. 2005: 709- 716
[8]刘红梅,一种小波变换域的自适应视频水印算法[J],电子学报,2001, 29(12): 1656-1660
[9]张建伟,鲍政,王顺凤,图像小波域分块奇异值分解的自适应水印方案,中国图像图形学报,Journal of Image and Graphics,2007, 12(05): 811-818
[10] Liu, R. and Tan, T. An SVD-Based Watermarking Scheme for Protecting Rightful Ownership[J]. IEEE Transactions on Multimedia(S1070-986X). 2002, 4(1): 121-128.
[11] Chandra, D. V. S. Digital Image Watermarking Using Singular Value Decomposition[C]. In Proceedings of 45th Midwest Symposium on Circuits and Systems (MWSCAS02): 2002. p264-267
[12] Quan, L. and Qingsong, A. A Combination of DCT-Based and SVD-Based Watermarking Scheme[C]. In Proceedings of 7th International Conference on Signal Processing (ICSP04), 2004(8):873-876.
[13] Ganic, E. and Eskicioglu, A. M. Robust Embedding of Visual Watermarks Using DWT-SVD.2005. http://www.sci.brooklyn.cuny.edu/~eskicioglu/papers/JEI2005.pdf
[14] Sverdlov, A., Dexter, S., Eskicioglu, A. M. Robust DCT SVD Domain Image Watermarking for Copyright Protection: Embedding Data in All Frequency[C]. In 13th European Signal Processing Conference (EUSIPCO '05) , Antalya, Turkey:2005.
[15]程云鹏,矩阵论,西北工业大学出版社,1989
[16] P Drineas, MW Mahoney , A randomized algorithm for a tensor-based generalization of the singular value decomposition,Linear Algebra and Its Applications, 2007,420(2-3):553-571
[17] K . Ting , P . Fung , C . Chang , F . Chan,Automatic correction of artifact from single-trial event-related potentials by blind source separation using second order statistics only,Medical Engineering & Physics , 2007,28(8):780– 794
[18]游荣义,陈忠,一种基于ICA的盲信号分离快速算法,电子学报,2004,32(4):669-672
[19]马建仓,牛奕龙,陈海洋,盲信号处理,北京:国防工业出版社,2006
[20]程云鹏,矩阵论,西北工业大学出版社,1989
[21] P Drineas, MW Mahoney , A randomized algorithm for a tensor-based generalization of the singular value decomposition,Linear Algebra and Its Applications, 2007,420(2-3):553-571
[22] V Aslantas,A singular-value decomposition-based image watermarking using genetic algorithm , AEUE-International Journal of Electronics and Communications, 2008,62(5):386-394
[23] GCW Ting, BM Goi, SH Heng,Attack on a semi-blind watermarking scheme based on singular value decomposition,Computer Standards & Interfaces, 2009,31(2):523-525
[24] F Truchetet, O Laligant,A review on industrial applications of wavelet and multiresolution based signal-image processing,Journal Electronic Imaging, 2008:23-25
[25]姜学思,韩秀莉,刘宁,奇异值数字水印算法的数学基础,陕西教育:高教版, 2008
[26]杨文杰,孙小军,胡明昊,基于特征点的抗几何攻击水印算法,计算机工程与应用,2008(19):23-28
[27] Ruanaidh Jó, Pun T. Rotation, Scale, and Translation Invariant Digital Image Watermarking[J]. Signal Processing, 1998, 66(3):303-317.
[28] Kim B S, Choi J G, Park C H, et al. Robust Digital Image Watermarking Method Against Geometrical Attacks[J]. Real-timeImaging, 2003, 9(2): 139-149.
[29]李海峰,宋巍巍,王树勋,基于Contourlet变换的鲁棒性图像水印算法,通信学报,2006,27(4):119-127
[30] Zhao Xu, Ke Wang and Xiao-hua Qiao, A Novel Watermarking Scheme inContourlet Domain Based on Independent Component Analysis, Proceedings of the 2006 International Conference on Intelligent Information Hiding and Multimedia Signal Processing IEEE, 2006:59-62
[31] Baaziz Nadia·Adaptive watermarking schemes based on a redundant Contourlet transform [C],Proceedings of the International Conference on Image Processing, Genoa, 2005(1): 221-224
[32] Jayalakshmi M., S. N. Merchant, Uday B. Desai,DigitalWatermarking in Contourlet Domain,The 18th International Conference on Pattern Recognition 2006(3):861-864
[33] Hu M K. Visual Pattern Recognition by Moment Invariants [J] . IRE Trans. Info. Theory, 1962, 8 (2):179 187.
[34]高珍,台莉春,张志浩,数字水印对抗几何攻击方法的研究,计算机工程,2006,32(5):156-159
[35] M N Do, M Vetterl.i Contourlets[A]. J Stoeckler, G V Welland. Beyond Wavelets [M]. Academic Press, 2002.
[36]李海峰,宋巍巍,王树勋,基于Contourlet变换的鲁棒性图像水印算法,通信学报,2006,27(4):119-127
[37] Zhao Xu, Ke Wang and Xiao-hua Qiao, A Novel Watermarking Scheme in Contourlet Domain Based on Independent Component Analysis, Proceedings of the 2006 International Conference on Intelligent Information Hiding and Multimedia Signal Processing IEEE, 2006:59-62
[38] Baaziz Nadia·Adaptive watermarking schemes based on a redundant Contourlet transform [C],Proceedings of the International Conference on Image Processing, Genoa, 2005(1): 221-224
[39]梁栋,殷兵,于梅,李新华,王年,基于非抽样Contourlet变换的自适应阈值图像增强算法,电子学报,2008,36(3):238-242
[40]侯正信,任亮,郭旭静,一种新的全相位Contourlet离散变换在图像去噪中的应用,中国图象图形学报,2008,13(5):124-129
[41] FLOM L, SAFIR A. Iris recognition system[P]. Washington: Patent and Trademark Office, 1987.
[42] Jianzhen Wi andJianyig Xie, adaptive image watermarking scheme based on HVS and fuzzy clustering theory[C], IEEE Int. Conf. Neural Networks & Signal Processing, December 14-17, 2003: 493-1496
[43] Said E. El-Khamy, Noha A. El-Yamny, A New Perceptual Image Watermarking Technique Based on Adaptive Fuzzy Clustering[C], 21" National Radio ScienceConference (NRSC2004), (NTI) March 16-18, 2004, C30: 1-8
[44] Chip-Hong Chang; Zhi Ye; Mingyan Zhang; Fuzzy-ART based adaptive digital watermarking scheme[J], Circuits and Systems for Video Technology, IEEE Transactions, Jan. 2005, 15(1): 65-81
[45]刘向丽,寇卫东,王志国,一种新的非对称数字水印算法,西安电子科技大学学报,2008,35(01):129-132
[46] Sudhir Rao, Weifeng Liu, Principe, J.C. de Medeiros Martins A., Information Theoretic Mean Shift Algorithm, Machine Learning for Signal Processing, 2006[C]. Proceedings of the 2006 16th IEEE Signal Processing Society Workshop, Sept. 2006: 155-160.
[47] Wenbing Tao, Hai Jin, Yimin Zhang, Color Image Segmentation Based on Mean Shift and Normalized Cuts[J], IEEE Transactions on Systems Man and Cybernetics Part B-Cybernetics.2007,37(5):1382 - 1389
[48] Collins,Robert T.Mean-shift blob tracking through scale space[M].United States: Institute of Electrical and Electronics Engineers Computer Society,2003:234-240.
[49] Fukunaga, K. Hostetler L.,The estimation of the gradient of a density function, with applications in pattern recognition,IEEE Transactions on Information Theory,1975,21(1):32- 40
[50] MN Do, MC Vetterli,Beyond Wavelets,Academic Press,2002:83-107
[51]文志强,蔡自兴,Mean Shift算法的收敛性分析[J] Journal of Software, 2007, 18(2):205?212
[52] Baaziz N.,Adaptive watermarking schemes based on a redundant Contourlet transform,IEEE International Conference on Image Processing,2005(1):221-224
[2]范科峰,莫玮,曹山,赵新华,裴庆祺,数字版权管理技术及应用研究进展,电子学报,2007,35(6):1139-1147
[3]和红杰,陈帆,自嵌入水印算法的安全性分析,电子学报,2007,35(3):557-562
[4]王炳锡,陈琦,邓峰森,数字水印技术,西安:西安电子科技大学出版社, 2003
[5]孙圣和,陆哲明,牛夏牧,数字水印技术及应用,北京:科学出版社, 2004
[6]钮心忻,信息隐藏与数字水印,北京:北京邮电大学出版社,2004
[7] Potdar. V. M., Han. S, Chang. E., A survey of digital image watermarking techniques, 2005 3rd IEEE International Conference on Industrial Informatics, 2005. INDIN '05. 10-12 Aug. 2005: 709- 716
[8]刘红梅,一种小波变换域的自适应视频水印算法[J],电子学报,2001, 29(12): 1656-1660
[9]张建伟,鲍政,王顺凤,图像小波域分块奇异值分解的自适应水印方案,中国图像图形学报,Journal of Image and Graphics,2007, 12(05): 811-818
[10] Liu, R. and Tan, T. An SVD-Based Watermarking Scheme for Protecting Rightful Ownership[J]. IEEE Transactions on Multimedia(S1070-986X). 2002, 4(1): 121-128.
[11] Chandra, D. V. S. Digital Image Watermarking Using Singular Value Decomposition[C]. In Proceedings of 45th Midwest Symposium on Circuits and Systems (MWSCAS02): 2002. p264-267
[12] Quan, L. and Qingsong, A. A Combination of DCT-Based and SVD-Based Watermarking Scheme[C]. In Proceedings of 7th International Conference on Signal Processing (ICSP04), 2004(8):873-876.
[13] Ganic, E. and Eskicioglu, A. M. Robust Embedding of Visual Watermarks Using DWT-SVD.2005. http://www.sci.brooklyn.cuny.edu/~eskicioglu/papers/JEI2005.pdf
[14] Sverdlov, A., Dexter, S., Eskicioglu, A. M. Robust DCT SVD Domain Image Watermarking for Copyright Protection: Embedding Data in All Frequency[C]. In 13th European Signal Processing Conference (EUSIPCO '05) , Antalya, Turkey:2005.
[15]程云鹏,矩阵论,西北工业大学出版社,1989
[16] P Drineas, MW Mahoney , A randomized algorithm for a tensor-based generalization of the singular value decomposition,Linear Algebra and Its Applications, 2007,420(2-3):553-571
[17] K . Ting , P . Fung , C . Chang , F . Chan,Automatic correction of artifact from single-trial event-related potentials by blind source separation using second order statistics only,Medical Engineering & Physics , 2007,28(8):780– 794
[18]游荣义,陈忠,一种基于ICA的盲信号分离快速算法,电子学报,2004,32(4):669-672
[19]马建仓,牛奕龙,陈海洋,盲信号处理,北京:国防工业出版社,2006
[20]程云鹏,矩阵论,西北工业大学出版社,1989
[21] P Drineas, MW Mahoney , A randomized algorithm for a tensor-based generalization of the singular value decomposition,Linear Algebra and Its Applications, 2007,420(2-3):553-571
[22] V Aslantas,A singular-value decomposition-based image watermarking using genetic algorithm , AEUE-International Journal of Electronics and Communications, 2008,62(5):386-394
[23] GCW Ting, BM Goi, SH Heng,Attack on a semi-blind watermarking scheme based on singular value decomposition,Computer Standards & Interfaces, 2009,31(2):523-525
[24] F Truchetet, O Laligant,A review on industrial applications of wavelet and multiresolution based signal-image processing,Journal Electronic Imaging, 2008:23-25
[25]姜学思,韩秀莉,刘宁,奇异值数字水印算法的数学基础,陕西教育:高教版, 2008
[26]杨文杰,孙小军,胡明昊,基于特征点的抗几何攻击水印算法,计算机工程与应用,2008(19):23-28
[27] Ruanaidh Jó, Pun T. Rotation, Scale, and Translation Invariant Digital Image Watermarking[J]. Signal Processing, 1998, 66(3):303-317.
[28] Kim B S, Choi J G, Park C H, et al. Robust Digital Image Watermarking Method Against Geometrical Attacks[J]. Real-timeImaging, 2003, 9(2): 139-149.
[29]李海峰,宋巍巍,王树勋,基于Contourlet变换的鲁棒性图像水印算法,通信学报,2006,27(4):119-127
[30] Zhao Xu, Ke Wang and Xiao-hua Qiao, A Novel Watermarking Scheme inContourlet Domain Based on Independent Component Analysis, Proceedings of the 2006 International Conference on Intelligent Information Hiding and Multimedia Signal Processing IEEE, 2006:59-62
[31] Baaziz Nadia·Adaptive watermarking schemes based on a redundant Contourlet transform [C],Proceedings of the International Conference on Image Processing, Genoa, 2005(1): 221-224
[32] Jayalakshmi M., S. N. Merchant, Uday B. Desai,DigitalWatermarking in Contourlet Domain,The 18th International Conference on Pattern Recognition 2006(3):861-864
[33] Hu M K. Visual Pattern Recognition by Moment Invariants [J] . IRE Trans. Info. Theory, 1962, 8 (2):179 187.
[34]高珍,台莉春,张志浩,数字水印对抗几何攻击方法的研究,计算机工程,2006,32(5):156-159
[35] M N Do, M Vetterl.i Contourlets[A]. J Stoeckler, G V Welland. Beyond Wavelets [M]. Academic Press, 2002.
[36]李海峰,宋巍巍,王树勋,基于Contourlet变换的鲁棒性图像水印算法,通信学报,2006,27(4):119-127
[37] Zhao Xu, Ke Wang and Xiao-hua Qiao, A Novel Watermarking Scheme in Contourlet Domain Based on Independent Component Analysis, Proceedings of the 2006 International Conference on Intelligent Information Hiding and Multimedia Signal Processing IEEE, 2006:59-62
[38] Baaziz Nadia·Adaptive watermarking schemes based on a redundant Contourlet transform [C],Proceedings of the International Conference on Image Processing, Genoa, 2005(1): 221-224
[39]梁栋,殷兵,于梅,李新华,王年,基于非抽样Contourlet变换的自适应阈值图像增强算法,电子学报,2008,36(3):238-242
[40]侯正信,任亮,郭旭静,一种新的全相位Contourlet离散变换在图像去噪中的应用,中国图象图形学报,2008,13(5):124-129
[41] FLOM L, SAFIR A. Iris recognition system[P]. Washington: Patent and Trademark Office, 1987.
[42] Jianzhen Wi andJianyig Xie, adaptive image watermarking scheme based on HVS and fuzzy clustering theory[C], IEEE Int. Conf. Neural Networks & Signal Processing, December 14-17, 2003: 493-1496
[43] Said E. El-Khamy, Noha A. El-Yamny, A New Perceptual Image Watermarking Technique Based on Adaptive Fuzzy Clustering[C], 21" National Radio ScienceConference (NRSC2004), (NTI) March 16-18, 2004, C30: 1-8
[44] Chip-Hong Chang; Zhi Ye; Mingyan Zhang; Fuzzy-ART based adaptive digital watermarking scheme[J], Circuits and Systems for Video Technology, IEEE Transactions, Jan. 2005, 15(1): 65-81
[45]刘向丽,寇卫东,王志国,一种新的非对称数字水印算法,西安电子科技大学学报,2008,35(01):129-132
[46] Sudhir Rao, Weifeng Liu, Principe, J.C. de Medeiros Martins A., Information Theoretic Mean Shift Algorithm, Machine Learning for Signal Processing, 2006[C]. Proceedings of the 2006 16th IEEE Signal Processing Society Workshop, Sept. 2006: 155-160.
[47] Wenbing Tao, Hai Jin, Yimin Zhang, Color Image Segmentation Based on Mean Shift and Normalized Cuts[J], IEEE Transactions on Systems Man and Cybernetics Part B-Cybernetics.2007,37(5):1382 - 1389
[48] Collins,Robert T.Mean-shift blob tracking through scale space[M].United States: Institute of Electrical and Electronics Engineers Computer Society,2003:234-240.
[49] Fukunaga, K. Hostetler L.,The estimation of the gradient of a density function, with applications in pattern recognition,IEEE Transactions on Information Theory,1975,21(1):32- 40
[50] MN Do, MC Vetterli,Beyond Wavelets,Academic Press,2002:83-107
[51]文志强,蔡自兴,Mean Shift算法的收敛性分析[J] Journal of Software, 2007, 18(2):205?212
[52] Baaziz N.,Adaptive watermarking schemes based on a redundant Contourlet transform,IEEE International Conference on Image Processing,2005(1):221-224