数字图像压缩历史的分析与检测
|
摘要
随着数字图像的广泛应用,其真实性的鉴别成为了一个研究的热点问题,图像处理软件越来越多,技术也越来越复杂精细,使得更多的人能够轻易地对图像进行篡改。篡改手段的多样化使得数字图像取证工作更加艰难,目前已有很多学者针对图像篡改过程如复制-粘贴、模糊、尺度变换等操作提出了多种方法对其进行检测,从而对数字图像的真实性做出鉴别。
数字图像压缩历史检测是数字图像取证的一个重要的组成部分,通过判断图像之前是否经过压缩,可对数字图像的真实性作出一个初步的判断,可作为数字图像取证的辅助决策。本文在这一研究背景下对数字图像压缩历史的分析与检测进行了深入的研究,主要的研究成果如下:
(1)对数字图像压缩历史检测进行了全面系统的介绍,包括研究背景和研究现状等,并对现有的几种检测技术包括无损图像的压缩历史检测和有损压缩图像的重压缩检测进行详细的介绍和分析,更重要的是将数字图像压缩历史的分析与检测融入到数字图像取证研究中,分析了其在数字图像取证领域的应用包括检测合成篡改、隐密分析和图像来源等。
(2)在深入研究分析JPEG压缩原理及其对数字图像JPEG系数的影响的基础上,提出了基于图像信息损失量的无损图像压缩历史检测算法。本文通过将待测图像剪切四行四列来估计原始图像,并提出使用图像信息损失量来度量图像的信息损失及与原始图像的相似程度,通过提取统计特征,利用Fisher线性分类器和阈值分割方法对图像进行压缩历史检测。实验结果表明,该算法受图像内容影响很小,而且能够不经训练测试就能够对单幅图像进行准确判断。
(3)针对JPEG压缩图像重压缩检测方面,在以往重压缩检测算法的基础上,提取了能够反映重压缩特性的直方图特征、DFT特征、Benford特征和Markov特征,在实际应用中,图像的原始压缩质量因数未知,因此不同于以往的文献中采用的一对一的测试方法,本文分别对这几种特征进行原始压缩质量因数不同的混合训练和测试,对其实验结果进行分析和比较。
The authentication of digital images is attracting more and more attention from researchers. Along with the wide use of image processing softwares, it becomes more easily for people to tamper the image content which makes image forensics much harder. A lot of methods have been proposed to detect tampering process such as copy-paste, blurring, scale transformation and so on.
As an important part of digital image forensics, digital image compression history detection can be used to make a first judgment of the authenticity of images. Under this circumstance, this paper focuses on the research of compression history analysis and detection of digital images. The main contribution of this thesis includes:
(1) The paper gives a comprehensive introduction to the compression history detection of digital images including the researching background and existing technical literatures on the compression history detection of lossless images(BMP) and loss compressed images(JPEG). Furthermore, the paper applies the analysis and detection of image compression history to image forensics such as copy-paste detection, steganalysis and image source determination.
(2) The approach of deting compression history of BMP images based on image information loss is proposed in this paper. Through analyzing the effect of JPEG compression to JPEG coefficients, statistical features reflecting image information loss difference of the detected image and the estimated image can be extracted. Thus using Fisher linear classifier and threshold division the compression history of BMP images are detected in this paper. Experiments show that the detection accuracy of the proposed approach is high, and not affected by image contents compared to Benford methods. It can also make an accurate judge of the compression history of a single BMP image without training.
(3) The paper exacted features which can reflect the double compression effect including histogram features, DFT features, Benford features and Markov features of JPEG coefficients based on previous double compression detection algorithms. The original compression quality of images is usually unkown in practical use, therefore, different from exisiting literatures which use one-to-one test methods, this paper did some experiments on mixed image database with different original compress quality and comparade the results using different features.
数字图像压缩历史检测是数字图像取证的一个重要的组成部分,通过判断图像之前是否经过压缩,可对数字图像的真实性作出一个初步的判断,可作为数字图像取证的辅助决策。本文在这一研究背景下对数字图像压缩历史的分析与检测进行了深入的研究,主要的研究成果如下:
(1)对数字图像压缩历史检测进行了全面系统的介绍,包括研究背景和研究现状等,并对现有的几种检测技术包括无损图像的压缩历史检测和有损压缩图像的重压缩检测进行详细的介绍和分析,更重要的是将数字图像压缩历史的分析与检测融入到数字图像取证研究中,分析了其在数字图像取证领域的应用包括检测合成篡改、隐密分析和图像来源等。
(2)在深入研究分析JPEG压缩原理及其对数字图像JPEG系数的影响的基础上,提出了基于图像信息损失量的无损图像压缩历史检测算法。本文通过将待测图像剪切四行四列来估计原始图像,并提出使用图像信息损失量来度量图像的信息损失及与原始图像的相似程度,通过提取统计特征,利用Fisher线性分类器和阈值分割方法对图像进行压缩历史检测。实验结果表明,该算法受图像内容影响很小,而且能够不经训练测试就能够对单幅图像进行准确判断。
(3)针对JPEG压缩图像重压缩检测方面,在以往重压缩检测算法的基础上,提取了能够反映重压缩特性的直方图特征、DFT特征、Benford特征和Markov特征,在实际应用中,图像的原始压缩质量因数未知,因此不同于以往的文献中采用的一对一的测试方法,本文分别对这几种特征进行原始压缩质量因数不同的混合训练和测试,对其实验结果进行分析和比较。
The authentication of digital images is attracting more and more attention from researchers. Along with the wide use of image processing softwares, it becomes more easily for people to tamper the image content which makes image forensics much harder. A lot of methods have been proposed to detect tampering process such as copy-paste, blurring, scale transformation and so on.
As an important part of digital image forensics, digital image compression history detection can be used to make a first judgment of the authenticity of images. Under this circumstance, this paper focuses on the research of compression history analysis and detection of digital images. The main contribution of this thesis includes:
(1) The paper gives a comprehensive introduction to the compression history detection of digital images including the researching background and existing technical literatures on the compression history detection of lossless images(BMP) and loss compressed images(JPEG). Furthermore, the paper applies the analysis and detection of image compression history to image forensics such as copy-paste detection, steganalysis and image source determination.
(2) The approach of deting compression history of BMP images based on image information loss is proposed in this paper. Through analyzing the effect of JPEG compression to JPEG coefficients, statistical features reflecting image information loss difference of the detected image and the estimated image can be extracted. Thus using Fisher linear classifier and threshold division the compression history of BMP images are detected in this paper. Experiments show that the detection accuracy of the proposed approach is high, and not affected by image contents compared to Benford methods. It can also make an accurate judge of the compression history of a single BMP image without training.
(3) The paper exacted features which can reflect the double compression effect including histogram features, DFT features, Benford features and Markov features of JPEG coefficients based on previous double compression detection algorithms. The original compression quality of images is usually unkown in practical use, therefore, different from exisiting literatures which use one-to-one test methods, this paper did some experiments on mixed image database with different original compress quality and comparade the results using different features.
引文
[1]http://www.cs.dartmouth.edu/farid/research/digitaltampering/
[2]Hany Farid.Creating and Detecting Doctored and Virtual Images:implications to The Child Pornography Prevention Act[EB/OL].http://www.ists.dart mouth.edu/library /tr-2004- 518.pdf
[3]周琳娜.数字图像盲取证技术研究:(博士学位论文).北京:北京邮电大学,2007.
[4]Fan Z,de Queiroz R.Maximum likelihood estimation of jpeg quantization table in the identification of bitmap compression history.In:Proceedings of ICIP' 00,10-13 Sept.2000,1:948-951.
[5]Z.Fan and R.L.Queiroz.Identification of bitmap compression history:JPEG detection and quantizer estimation.IEEE Transaction on Image Processing,vol.12,no.2,Feb.2003.
[6]Ramesh(Neelsh) Neelamani,Ricardo de Queiroz,Zhigang Fan et al.JPEG Compression History Estimationfor Color Images.IEEE Transaction on Image Processing,VOL.15,NO.6,JUNE 2006.
[7]Popescu A.Statistical Tools for Digital Image Forensics.Ph.D.thesis,Department of Computer Science,Dartmouth College,Hanover,NH,2005.
[8]A C Popescu,H Farid.Statistical Tools for Digital Image Forensics.In:Proc of Information Hiding 6th International Workshop.Berlin Heidelberg:Springer Press,2004.128-147.
[9]J Fridrich,J Lukas.Estimation of primary quantization matrix in double compressed JPEG images.Digital Forensics Research Workshop.Cleveland,Ohio,2003
[10]T Pevny,J Fridrich.Detection of Double-Compression in JPEG Images for Applications in Steganography.IEEE Transactions on Information Forensics and Security,2008 3(2):247~258
[11]D D Fu,Y Q Shi,W Su.A generalized Benford' s law for JPEG coefficients and its applications in image forensics.In:Proc of SPIE Security,Steganography,and Watermarking of Multimedia Contents Ⅸ.Bellingham,Washington:SPIE Press,2007.1L,1~11.
[12]Chunhua Chen,Yun Q.Shi,and Wei Su,A Machine Learning Based Scheme for Double JPEG Compression Detection,Pattern Recognition,2008.ICPR 2008.19th International Conference on.Page(s):1-4.Dec.2008.
[13]W.-Q.Luo,Z.-H.Qu,J.-Wu.Huang,and G.-P.Qiu,"A novel method for detecting cropped and recompressed image block," in Proc.IEEE ICASSP' 07,Hawaii,2007.
[14]Zhenhua Qu,Weiqi LuG,Jiwu Huang,A convolutive mixing model for shifted double JPEG compression with application to passive image authentication,in Proc.IEEE ICASSP' 08,2008.
[15]He J F,Lin Z C,Wang L F,Tang X O.Detecting doctored JPEG images via DCT coefficient analysis.In:Proceedings of 9th European Conference on Computer Vision,Graz,Austria,2006.423-435.
[16]戴蒙,林家骏,毛家发.JPEG二次压缩的分析与检测.中国图象图形学报,2006,11(11):(1619-1622).
[17]岳红原.基于统计的数字图像篡改检测方法[D].上海:上海师范大学,2006.
[18]于雪燕,一种基于双JPEG压缩的数字图像篡改的检测方法[D].上海:上海师范大学,2007.
[19]李晟,张新鹏.利用JPEG压缩特性的合成图像检测[J].应用科学学报,2008,(3).
[20]Fridrich J.,Goljan M.,and Hogea D.:Attacking the OutGuess,Proc.Multimedia and Security,Workshop at ACM Multimedia,Juan-les-Pins,France,December 6,2002.
[21]Fridrich J.,Goljan M.,and Hogea D.:Steganalysis of JPEG Images:Breaking the F5 Algorithm,Proc.5th International Workshop on Information Hiding,Noordwijkerhout,The Nederlands,October 7-9 2002,pp.310-323.
[22]G K Wallace.The JPEG Still Picture Compression Standard.IEEE Transactions on Consumer Electronics,1992 38(1):ⅹⅴⅲ~ⅹⅹⅹⅸ.
[23]A.Westfeld and A.Pfitzmann.Attacks on Steganographic Systems.IHW 99,Dresden,Germany,1999.
[24]Fridrich J.,Goljan M.,and Hogea D.:"Attacking the OutGuess",Proc.Multimedia and Security,Workshop at ACM Multimedia,Juan-les-Pins,France,December 6,2002.
[25]Fridrich.Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganographic Schemes,Proc.6th Information Hiding Workshop,Toronto,Canada,May 23-25,2004.
[26]Fridrich J,Goljan M,Du R.Steganalysis based on jpeg compatibility,In:Proceedings of SPIE Electronic Imaging,Multimedia Systemsand Applications,2001,4518(1):275 - 280.
[27]Shuiming Ye,Detecting digital image forgeries by measuring inconsistencies of blocking artifact,Multimedia and Expo,2007 IEEE International Conference on,July 2007.
[28]Jpegsr6[EB/OL].ftp://ftp.simtel.net/pub/simtelnet/msdos/graphics/jpegsr6.zip
[29]Farid Hany.Digital image ballistics from JPEG quantization.Technical Report TR2006-583,Department of Computer Science,Dartmouth College;2006.
[30]Jesse D.Kornblum,Using JPEG quantization tables to identify imagery processed by software,Digital Investigations,Elsevier,2008.
[31] T Pevny, J Fridrich. Determining the stego algorithm for JPEG images. In: Proc of Information Security. Stevenage, Herts: IEE Press, 2006. 77-86.
[32] Tian-Tsong Ng, Shih-Fu Chang, Qibin Sun. Blind Detection of Photomontage Using Higer Order Statistics. In IEEE International Symposium on Circuits and Systems(ISCAS), Vancouver, Canada, May 2004.
[33] J. -Y. Lee and H. W. Park, "A rate control algorithm for DCT-based video coding using simple rate estimation and linear source model" , IEEE Transactions on Circuits and Systems for Video Technology, vol. 15, no. 9, pp. 1077 - 1085, 2005.
[34] Kai San Choi, Edmund Y. Lam, Kenneth K. Y. Wong, "Source Camera Identification by JPEG Compression Statistics for Image Forensics" , IEEE, 2006.
[35] Fridrich, J. , Goljan, M., Hogea, D. , and Soukal, D. : Quantitative Steganalysis:Estimating Secret Message Length. ACM Multimedia Systems Journal. Special issue on Multimedia Security, Vol. 9(3) (2003) 288-302.
[36] C.-C. Chang, C.-J. Lin. LIBSVM: a library for support vector machines.http://www.csie.ntu.edu.tw /～cjlin. 2007,6.
[37] C.-W. Hsu, C.-C. Chang, C.-J. Lin. A Practical Guide to Support Vector Classification. http://www. csie. ntu. edu. tw/"cjlin/papers/guide/guide. pdf, 2007,7.
[2]Hany Farid.Creating and Detecting Doctored and Virtual Images:implications to The Child Pornography Prevention Act[EB/OL].http://www.ists.dart mouth.edu/library /tr-2004- 518.pdf
[3]周琳娜.数字图像盲取证技术研究:(博士学位论文).北京:北京邮电大学,2007.
[4]Fan Z,de Queiroz R.Maximum likelihood estimation of jpeg quantization table in the identification of bitmap compression history.In:Proceedings of ICIP' 00,10-13 Sept.2000,1:948-951.
[5]Z.Fan and R.L.Queiroz.Identification of bitmap compression history:JPEG detection and quantizer estimation.IEEE Transaction on Image Processing,vol.12,no.2,Feb.2003.
[6]Ramesh(Neelsh) Neelamani,Ricardo de Queiroz,Zhigang Fan et al.JPEG Compression History Estimationfor Color Images.IEEE Transaction on Image Processing,VOL.15,NO.6,JUNE 2006.
[7]Popescu A.Statistical Tools for Digital Image Forensics.Ph.D.thesis,Department of Computer Science,Dartmouth College,Hanover,NH,2005.
[8]A C Popescu,H Farid.Statistical Tools for Digital Image Forensics.In:Proc of Information Hiding 6th International Workshop.Berlin Heidelberg:Springer Press,2004.128-147.
[9]J Fridrich,J Lukas.Estimation of primary quantization matrix in double compressed JPEG images.Digital Forensics Research Workshop.Cleveland,Ohio,2003
[10]T Pevny,J Fridrich.Detection of Double-Compression in JPEG Images for Applications in Steganography.IEEE Transactions on Information Forensics and Security,2008 3(2):247~258
[11]D D Fu,Y Q Shi,W Su.A generalized Benford' s law for JPEG coefficients and its applications in image forensics.In:Proc of SPIE Security,Steganography,and Watermarking of Multimedia Contents Ⅸ.Bellingham,Washington:SPIE Press,2007.1L,1~11.
[12]Chunhua Chen,Yun Q.Shi,and Wei Su,A Machine Learning Based Scheme for Double JPEG Compression Detection,Pattern Recognition,2008.ICPR 2008.19th International Conference on.Page(s):1-4.Dec.2008.
[13]W.-Q.Luo,Z.-H.Qu,J.-Wu.Huang,and G.-P.Qiu,"A novel method for detecting cropped and recompressed image block," in Proc.IEEE ICASSP' 07,Hawaii,2007.
[14]Zhenhua Qu,Weiqi LuG,Jiwu Huang,A convolutive mixing model for shifted double JPEG compression with application to passive image authentication,in Proc.IEEE ICASSP' 08,2008.
[15]He J F,Lin Z C,Wang L F,Tang X O.Detecting doctored JPEG images via DCT coefficient analysis.In:Proceedings of 9th European Conference on Computer Vision,Graz,Austria,2006.423-435.
[16]戴蒙,林家骏,毛家发.JPEG二次压缩的分析与检测.中国图象图形学报,2006,11(11):(1619-1622).
[17]岳红原.基于统计的数字图像篡改检测方法[D].上海:上海师范大学,2006.
[18]于雪燕,一种基于双JPEG压缩的数字图像篡改的检测方法[D].上海:上海师范大学,2007.
[19]李晟,张新鹏.利用JPEG压缩特性的合成图像检测[J].应用科学学报,2008,(3).
[20]Fridrich J.,Goljan M.,and Hogea D.:Attacking the OutGuess,Proc.Multimedia and Security,Workshop at ACM Multimedia,Juan-les-Pins,France,December 6,2002.
[21]Fridrich J.,Goljan M.,and Hogea D.:Steganalysis of JPEG Images:Breaking the F5 Algorithm,Proc.5th International Workshop on Information Hiding,Noordwijkerhout,The Nederlands,October 7-9 2002,pp.310-323.
[22]G K Wallace.The JPEG Still Picture Compression Standard.IEEE Transactions on Consumer Electronics,1992 38(1):ⅹⅴⅲ~ⅹⅹⅹⅸ.
[23]A.Westfeld and A.Pfitzmann.Attacks on Steganographic Systems.IHW 99,Dresden,Germany,1999.
[24]Fridrich J.,Goljan M.,and Hogea D.:"Attacking the OutGuess",Proc.Multimedia and Security,Workshop at ACM Multimedia,Juan-les-Pins,France,December 6,2002.
[25]Fridrich.Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganographic Schemes,Proc.6th Information Hiding Workshop,Toronto,Canada,May 23-25,2004.
[26]Fridrich J,Goljan M,Du R.Steganalysis based on jpeg compatibility,In:Proceedings of SPIE Electronic Imaging,Multimedia Systemsand Applications,2001,4518(1):275 - 280.
[27]Shuiming Ye,Detecting digital image forgeries by measuring inconsistencies of blocking artifact,Multimedia and Expo,2007 IEEE International Conference on,July 2007.
[28]Jpegsr6[EB/OL].ftp://ftp.simtel.net/pub/simtelnet/msdos/graphics/jpegsr6.zip
[29]Farid Hany.Digital image ballistics from JPEG quantization.Technical Report TR2006-583,Department of Computer Science,Dartmouth College;2006.
[30]Jesse D.Kornblum,Using JPEG quantization tables to identify imagery processed by software,Digital Investigations,Elsevier,2008.
[31] T Pevny, J Fridrich. Determining the stego algorithm for JPEG images. In: Proc of Information Security. Stevenage, Herts: IEE Press, 2006. 77-86.
[32] Tian-Tsong Ng, Shih-Fu Chang, Qibin Sun. Blind Detection of Photomontage Using Higer Order Statistics. In IEEE International Symposium on Circuits and Systems(ISCAS), Vancouver, Canada, May 2004.
[33] J. -Y. Lee and H. W. Park, "A rate control algorithm for DCT-based video coding using simple rate estimation and linear source model" , IEEE Transactions on Circuits and Systems for Video Technology, vol. 15, no. 9, pp. 1077 - 1085, 2005.
[34] Kai San Choi, Edmund Y. Lam, Kenneth K. Y. Wong, "Source Camera Identification by JPEG Compression Statistics for Image Forensics" , IEEE, 2006.
[35] Fridrich, J. , Goljan, M., Hogea, D. , and Soukal, D. : Quantitative Steganalysis:Estimating Secret Message Length. ACM Multimedia Systems Journal. Special issue on Multimedia Security, Vol. 9(3) (2003) 288-302.
[36] C.-C. Chang, C.-J. Lin. LIBSVM: a library for support vector machines.http://www.csie.ntu.edu.tw /～cjlin. 2007,6.
[37] C.-W. Hsu, C.-C. Chang, C.-J. Lin. A Practical Guide to Support Vector Classification. http://www. csie. ntu. edu. tw/"cjlin/papers/guide/guide. pdf, 2007,7.