数字图像水印算法的研究
摘要
随着信息技术和网络应用的飞速发展,版权认证和网络安全技术面临着极大的挑战。数字水印技术是解决这一问题的有效手段和方法,得到广泛的研究与应用。
本文介绍了国内外数字水印技术的研究背景、研究现状和发展趋势,分析了数字水印技术的原理、特征、分类、基本框架、典型算法、常见攻击和评价标准等。在此基础上,结合离散余弦变换、混沌理论、奇异值分解等理论提出了两种数字图像水印算法。
(1)DCT域的条码印刷水印算法。针对印刷品防伪技术,提出了一种将二值条码作为水印信息的印刷防伪算法,使水印经过一次打印扫描后具有鲁棒性,二次打印扫描后具有脆弱性。算法首先对水印信息进行混沌加密处理,实现了信息的初步隐藏；然后根据水印图像的大小对原始载体图像进行分块DCT变换,选取中低频系数并对其进行分组,通过分组中每两个系数的绝对值大小和符号的同异性对系数进行修改实现水印信息的嵌入,修改的幅度和范围较小,水印的不可见性好。根据嵌入水印后图像的分块DCT系数的正负性来提取水印,实现了盲提取。大量的仿真实验结果表明该算法具有良好的不可见性,符合印刷品防伪技术的要求。
(2)基于DCT与SVD的数字水印算法。首先在算法中引入了混沌技术,对原始水印图像进行加密,加密后的水印与原始水印的相似度很低,在不知道混沌初值的情况下,无法提取出水印。然后对载体图像进行分块DCT变换,将DCT系数按Zig-Zag排序并选取中低频系数组成系数矩阵,利用奇异值分解理论对系数矩阵进行变换,在其S分量上嵌入水印,有效地提高了水印的抗攻击能力。水印提取过程需要用到嵌入过程中奇异值分解的U和V分量。仿真实验验证了该算法不仅具有良好的不可见性,对于常见的攻击也具有较强的鲁棒性。
With the rapid development of the information technology and the network applications, copyright authentication and network security technology is facing great challenge. Digital watermarking technology is an effective method to solve the problem described above, and the technology was researched deeply and applied widely.
The research background, current situation and development trend of digital watermarking technology were described. The principles, characteristics, classification, basic framework, typical method, common assault and evaluation criteria of digital watermarking technology were analyzed. Based on the above basal theoretical research, combined with discrete cosine transform, chaos theory, singular value decomposition theory. Two digital image watermarking algorithms were proposed in this paper.
(1)The barcode printing watermarking algorithm based on DCT domain. The printed counterfeiting algorithm based on the printed counterfeiting technology was proposed, the binary barcode was regarded as the watermarking information in this algorithm. The advantage of this algorithm is that the watermarking is robust to the first PS (print and scan) and fragile to the second PS. At first, encrypting the watermarking information with chaotic method and the information can be initial hidden initially. Then, implementing the blocked DCT transform of the embedded watermarking image according to the size of the watermarking image. Selecting the low frequency coefficient and embedding the watermark information by modifying the coefficient of amplitude and plus-minus sign. During the watermarking extraction process, implementing the blocked DCT transform of the watermarked image at first, and then extracting the watermarking information according to the plus-minus sign of the DCT coefficients. The modifying extent is smaller and the algorithm has good invisibility. The watermarking was extracted according to plus-minus sign of the DCT coefficients, and the process is blind extraction. The simulation experiment results show that the algorithm has good invisibility, and the design of the algorithm accords with the requirement of the printed counterfeiting technology.
(2)The digital watermarking algorithm based on DCT and SVD. The chaos technology was pulled into the algorithm. Firstly, encrypting the original watermarking image with the chaos technology. The watermarking information can't be extracted if the chaos initial value is unknown, because the similarity of the encrypted watermarking and the original watermarking is low. Then, transforming the carrier image for block DCT, sorting the DCT coefficients according to Zig-Zag order and selecting the low frequency coefficients to constitute a coefficient matrix. Transforming the coefficient matrix with the Singular Value Decomposition method and embedding watermarking to the S component. The resist ability is improved in this algorithm. The SVD's U and V components produced during the embedding process should be used when extracting the watermarking. Simulation results show that the algorithm has good invisibility, while is better robust to the common attacks.
本文介绍了国内外数字水印技术的研究背景、研究现状和发展趋势,分析了数字水印技术的原理、特征、分类、基本框架、典型算法、常见攻击和评价标准等。在此基础上,结合离散余弦变换、混沌理论、奇异值分解等理论提出了两种数字图像水印算法。
(1)DCT域的条码印刷水印算法。针对印刷品防伪技术,提出了一种将二值条码作为水印信息的印刷防伪算法,使水印经过一次打印扫描后具有鲁棒性,二次打印扫描后具有脆弱性。算法首先对水印信息进行混沌加密处理,实现了信息的初步隐藏；然后根据水印图像的大小对原始载体图像进行分块DCT变换,选取中低频系数并对其进行分组,通过分组中每两个系数的绝对值大小和符号的同异性对系数进行修改实现水印信息的嵌入,修改的幅度和范围较小,水印的不可见性好。根据嵌入水印后图像的分块DCT系数的正负性来提取水印,实现了盲提取。大量的仿真实验结果表明该算法具有良好的不可见性,符合印刷品防伪技术的要求。
(2)基于DCT与SVD的数字水印算法。首先在算法中引入了混沌技术,对原始水印图像进行加密,加密后的水印与原始水印的相似度很低,在不知道混沌初值的情况下,无法提取出水印。然后对载体图像进行分块DCT变换,将DCT系数按Zig-Zag排序并选取中低频系数组成系数矩阵,利用奇异值分解理论对系数矩阵进行变换,在其S分量上嵌入水印,有效地提高了水印的抗攻击能力。水印提取过程需要用到嵌入过程中奇异值分解的U和V分量。仿真实验验证了该算法不仅具有良好的不可见性,对于常见的攻击也具有较强的鲁棒性。
With the rapid development of the information technology and the network applications, copyright authentication and network security technology is facing great challenge. Digital watermarking technology is an effective method to solve the problem described above, and the technology was researched deeply and applied widely.
The research background, current situation and development trend of digital watermarking technology were described. The principles, characteristics, classification, basic framework, typical method, common assault and evaluation criteria of digital watermarking technology were analyzed. Based on the above basal theoretical research, combined with discrete cosine transform, chaos theory, singular value decomposition theory. Two digital image watermarking algorithms were proposed in this paper.
(1)The barcode printing watermarking algorithm based on DCT domain. The printed counterfeiting algorithm based on the printed counterfeiting technology was proposed, the binary barcode was regarded as the watermarking information in this algorithm. The advantage of this algorithm is that the watermarking is robust to the first PS (print and scan) and fragile to the second PS. At first, encrypting the watermarking information with chaotic method and the information can be initial hidden initially. Then, implementing the blocked DCT transform of the embedded watermarking image according to the size of the watermarking image. Selecting the low frequency coefficient and embedding the watermark information by modifying the coefficient of amplitude and plus-minus sign. During the watermarking extraction process, implementing the blocked DCT transform of the watermarked image at first, and then extracting the watermarking information according to the plus-minus sign of the DCT coefficients. The modifying extent is smaller and the algorithm has good invisibility. The watermarking was extracted according to plus-minus sign of the DCT coefficients, and the process is blind extraction. The simulation experiment results show that the algorithm has good invisibility, and the design of the algorithm accords with the requirement of the printed counterfeiting technology.
(2)The digital watermarking algorithm based on DCT and SVD. The chaos technology was pulled into the algorithm. Firstly, encrypting the original watermarking image with the chaos technology. The watermarking information can't be extracted if the chaos initial value is unknown, because the similarity of the encrypted watermarking and the original watermarking is low. Then, transforming the carrier image for block DCT, sorting the DCT coefficients according to Zig-Zag order and selecting the low frequency coefficients to constitute a coefficient matrix. Transforming the coefficient matrix with the Singular Value Decomposition method and embedding watermarking to the S component. The resist ability is improved in this algorithm. The SVD's U and V components produced during the embedding process should be used when extracting the watermarking. Simulation results show that the algorithm has good invisibility, while is better robust to the common attacks.
引文
[1]冯运波,夏光升.信息安全技术发展现状[J].计算机安全,2002,21(9)：11-15
[2]Cox I.J,Matt L Miller. A Review of Watermarking and the Importance of Perceptual Modeling[C].SPIE Proceeding on Human Vision and Electronic Imaging,1997:92-99
[3]Su J.K, Hartung F, Girod B. Digital watermarking of text,image,and video documents[M]. Compute-rs&Graphics,1998,22(6):687-695
[4]Yu P-T,Tsai H.-H,Lin J.-S. Digital watermarking based on neural networks for color images[J].Signal Processing,2001,81(3):663-671
[5]Hartung F.,Girod B.Watermarking of uncompressed and compressed video[C].Signal Processing,1998,66(3):283-301
[6]Cvejic N.,Seppanen T. Spread spectrum audio watermarking using frequency hopping and atack characterization[C]. Signal Processing,2004,84(1):207-213
[7]王炳锡,陈琦,邓峰森.数字水印技术[M].西安：西安电子科技大学出社,2003,1-13
[8]杨义先,钮心忻.数字水印技术理论与应用[M].北京：高等教育出版社,2006,10-20
[9]王慧琴.数字图像处理[M].北京邮电大学出版社,2006,54-59
[10]钮心析.信息隐藏与数字水印[M].北京邮电大学出版社,2004,22-24
[11]孙圣和,陆哲明,牛夏牧等.数字水印技术及应用[M].北京：科学出版社,2004,31-33
[12]杜高峰.数字水印算法的研究与应用[D].西安：西安建筑科技大学,2007
[13]王丽娜,郭迟,李鹏著.信息隐藏技术实验教程[M].武汉大学出版社,2004,26-29
[14]钮心忻.信息隐藏与数字水印的研究及发展[J].计算机教育,2005年1月：29-31
[15]姜军,卓嘎,王朝霞.数字水印及其典型算法研究[J].四川师范大学学报(自然科学版),2009,上半年专刊
[16]蒋建国,宣曼,齐美彬.数字水印技术的研究现状及进展[J].计算机应用,2006,26(12)：60-62
[17]李春茹,王保保.数字水印稳健性攻击及相应策略[J].计算机仿真,2005,22(4)：82-85
[18]张小凡.基于小波变换的数字水印技术研究[D].大连：大连理工大学,2007
[19]Yiwei Wang,John Doherty,Robert van Dyck. A Wavelet-based Watermarking Algorithm for Ownership Verification of Digital Images [C].IEEE Transactions on Image Processing,2002,11 (2):77-78
[20]吴国民,庄越挺,吴飞等.基于视觉特征的视频水印技术[J].计算机辅助设计与图形学学报,2006,18(5)：715-721
[21]王向阳,杨红颖.DCT域自适应彩色图像二维数字水印算法研究[J].计算机辅助设计 与图形学学报,2004,16(2)：240-243
[22]杜高峰.数字水印算法的研究与应用[D].西安：西安建筑科技大学,2007
[23]王笋.基于混沌的数字水印技术研究[J].计算机应用,2007,25(11)：58-61
[24]张素芬.基于混沌加密的数字水印算法研究[D].山东：燕山大学,2005
[25]张宪海.数字水印技术在版权保护与内容认证中的应用研究[R].哈尔滨工程大学,2006：55-59
[26]Peter Toft. The Radon Transform-Theory and Implementation [D]. Denmark: Department of Mathematical Modelling,Technical University of Denmark.1996:5-104
[27]梁华庆,钮心忻,杨义先.基于数字水印的证件防伪技术[J].信号处理,2004,20(2)：143-146
[28]宋玉洁,刘瑞帧,谭铁牛等.数字水印技术在印刷品防伪中的应用[J].中国图象图形学报,2001,6(5)：450-454
[29]Kaushal Solanki, Upamanyu Madhow, BS Manjunath,etal. 'Print and Scan' Resilient Data Hiding in Images[J]. IEEE Transactions on Information Forensics and Security, 2006,1(4):464-478
[30]孙即祥.图像处理[M].北京：科学出版社,2004,64-69
[31]Kuashal Solanki,Upamanyu Madhow, Banglore S. Manjunath, and Shiv Chandrasekaran, Modeling the Print-Scan Process for Resilient Data Hiding, Proc. of SPIE:Securiyt, Steganography, and Watermarking of Multimedia Contents,2005:418-429
[32]冀芊茜.抗打印扫描数字水印算法研究及其在证件防伪中的应用[D].吉林：吉林大学,2007
[33]贺冲.数字图像水印算法的研究[D].西安：西安建筑科技大学,2009
[34]牛少彰.数字水印新算法的设计及鲁棒性分析[D].北京：北京邮电大学,2004
[35]冀芊茜.抗打印扫描数字水印算法研究及其在证件防伪中的应用[D].吉林：吉林大学,2007
[36]Liu R,Tan T.A SVD-based watermarking scheme for protecting rightful ownership[J]. IEEE Transactions on Multimedia,2002,4(1):121-128
[37]Zude Zhou,Bing Tang,Xinhua Liu.A Block-SVD Based Image Watermarking Method[C].Proceedings of the 6th World Congress on Intelligent Control and Au-tomation,2006:10347-10351
[38]付正立.抗几何攻击的数字图像水印研究[D].成都：电子科技大学,2009
[39]V. Senthil, R. Bhaskaran. Wavelet Based Digital Image Watermarking with Robustness against Geometric Attacks[C]. International Conference on Computational Intelligence and Multimedia Applications (ICCIMA),2007,157:89-93
[2]Cox I.J,Matt L Miller. A Review of Watermarking and the Importance of Perceptual Modeling[C].SPIE Proceeding on Human Vision and Electronic Imaging,1997:92-99
[3]Su J.K, Hartung F, Girod B. Digital watermarking of text,image,and video documents[M]. Compute-rs&Graphics,1998,22(6):687-695
[4]Yu P-T,Tsai H.-H,Lin J.-S. Digital watermarking based on neural networks for color images[J].Signal Processing,2001,81(3):663-671
[5]Hartung F.,Girod B.Watermarking of uncompressed and compressed video[C].Signal Processing,1998,66(3):283-301
[6]Cvejic N.,Seppanen T. Spread spectrum audio watermarking using frequency hopping and atack characterization[C]. Signal Processing,2004,84(1):207-213
[7]王炳锡,陈琦,邓峰森.数字水印技术[M].西安：西安电子科技大学出社,2003,1-13
[8]杨义先,钮心忻.数字水印技术理论与应用[M].北京：高等教育出版社,2006,10-20
[9]王慧琴.数字图像处理[M].北京邮电大学出版社,2006,54-59
[10]钮心析.信息隐藏与数字水印[M].北京邮电大学出版社,2004,22-24
[11]孙圣和,陆哲明,牛夏牧等.数字水印技术及应用[M].北京：科学出版社,2004,31-33
[12]杜高峰.数字水印算法的研究与应用[D].西安：西安建筑科技大学,2007
[13]王丽娜,郭迟,李鹏著.信息隐藏技术实验教程[M].武汉大学出版社,2004,26-29
[14]钮心忻.信息隐藏与数字水印的研究及发展[J].计算机教育,2005年1月：29-31
[15]姜军,卓嘎,王朝霞.数字水印及其典型算法研究[J].四川师范大学学报(自然科学版),2009,上半年专刊
[16]蒋建国,宣曼,齐美彬.数字水印技术的研究现状及进展[J].计算机应用,2006,26(12)：60-62
[17]李春茹,王保保.数字水印稳健性攻击及相应策略[J].计算机仿真,2005,22(4)：82-85
[18]张小凡.基于小波变换的数字水印技术研究[D].大连：大连理工大学,2007
[19]Yiwei Wang,John Doherty,Robert van Dyck. A Wavelet-based Watermarking Algorithm for Ownership Verification of Digital Images [C].IEEE Transactions on Image Processing,2002,11 (2):77-78
[20]吴国民,庄越挺,吴飞等.基于视觉特征的视频水印技术[J].计算机辅助设计与图形学学报,2006,18(5)：715-721
[21]王向阳,杨红颖.DCT域自适应彩色图像二维数字水印算法研究[J].计算机辅助设计 与图形学学报,2004,16(2)：240-243
[22]杜高峰.数字水印算法的研究与应用[D].西安：西安建筑科技大学,2007
[23]王笋.基于混沌的数字水印技术研究[J].计算机应用,2007,25(11)：58-61
[24]张素芬.基于混沌加密的数字水印算法研究[D].山东：燕山大学,2005
[25]张宪海.数字水印技术在版权保护与内容认证中的应用研究[R].哈尔滨工程大学,2006：55-59
[26]Peter Toft. The Radon Transform-Theory and Implementation [D]. Denmark: Department of Mathematical Modelling,Technical University of Denmark.1996:5-104
[27]梁华庆,钮心忻,杨义先.基于数字水印的证件防伪技术[J].信号处理,2004,20(2)：143-146
[28]宋玉洁,刘瑞帧,谭铁牛等.数字水印技术在印刷品防伪中的应用[J].中国图象图形学报,2001,6(5)：450-454
[29]Kaushal Solanki, Upamanyu Madhow, BS Manjunath,etal. 'Print and Scan' Resilient Data Hiding in Images[J]. IEEE Transactions on Information Forensics and Security, 2006,1(4):464-478
[30]孙即祥.图像处理[M].北京：科学出版社,2004,64-69
[31]Kuashal Solanki,Upamanyu Madhow, Banglore S. Manjunath, and Shiv Chandrasekaran, Modeling the Print-Scan Process for Resilient Data Hiding, Proc. of SPIE:Securiyt, Steganography, and Watermarking of Multimedia Contents,2005:418-429
[32]冀芊茜.抗打印扫描数字水印算法研究及其在证件防伪中的应用[D].吉林：吉林大学,2007
[33]贺冲.数字图像水印算法的研究[D].西安：西安建筑科技大学,2009
[34]牛少彰.数字水印新算法的设计及鲁棒性分析[D].北京：北京邮电大学,2004
[35]冀芊茜.抗打印扫描数字水印算法研究及其在证件防伪中的应用[D].吉林：吉林大学,2007
[36]Liu R,Tan T.A SVD-based watermarking scheme for protecting rightful ownership[J]. IEEE Transactions on Multimedia,2002,4(1):121-128
[37]Zude Zhou,Bing Tang,Xinhua Liu.A Block-SVD Based Image Watermarking Method[C].Proceedings of the 6th World Congress on Intelligent Control and Au-tomation,2006:10347-10351
[38]付正立.抗几何攻击的数字图像水印研究[D].成都：电子科技大学,2009
[39]V. Senthil, R. Bhaskaran. Wavelet Based Digital Image Watermarking with Robustness against Geometric Attacks[C]. International Conference on Computational Intelligence and Multimedia Applications (ICCIMA),2007,157:89-93