基于统计的数字图像篡改检测方法
|
摘要
由于功能强大的图像处理软件的广泛使用，数字图像越来越容易被篡改。现在，向图像中加入或从图像中移除某些重要的人或物，同时不留下明显的篡改痕迹已经成为可能。当数字相机和数字摄像机代替了传统的模拟相机和摄像机时，对数字图像的鉴别，图像内容的确认以及篡改的探测的需求就变得越来越迫切了。传统上，在鉴别图像真伪时，采用事先在数字图像中嵌入数字水印或签名的方法。但是，此方法存在着诸多的缺陷和不足。针对没有嵌入数字水印或签名的篡改图像，对其进行真伪的鉴定，成为新的研究方向。
一幅被改动过的数字图像可能用肉眼无法辨别真伪，但常常会因为篡改而留下一些不可见的线索。篡改会干扰图像一些潜在的统计特性。基于这样的设想，针对不同的图像篡改方式，本文提出了二种检测篡改图像的方法：
(1)基于双JPEG压缩统计特性的检测方法。这种方法是通过检测双JPEG压缩图像的图像块先前的量化系数来实现的。首先，定义了在两次压缩时，出现在特定位置上的DCT系数直方图的一些特性。然后，使用支持向量机来估计双压图像的初始的量化系数，有不同量化系数的图像块便是值得怀疑的区域。最后，基于大量不同的双压图像集，用一系列的实验来验证本方法的性能及可靠性。
(2)基于图像匹配的检测方法。此检测方法着重针对复制图像本身的一部分来遮盖另一重要部分的篡改类型(复制遮盖篡改)，针对不同尺寸大小的篡改图像提出了两种检测算法，并论证了检测算法的有效性和可靠性。这些方法对于复制区域被增强或修饰使之融入到背景之中，依然非常有效。
对于每一种方法，都论述了它的理论基础，并在实验结果中，展示了它在检测篡改图像时的有效性。
As a result of the widespread use of powerful image-processing software, it is increasingly easier to tamper digital images. For the time being, it has been possible to add or remove some important people or things to or from an image, leaving behind no obvious traces at the same time. Since digital cameras and digital vidicons replaced traditional analogue counterparts, it has become more and more urgent to identify digital images, to confirm image contents and detect their tampering. The traditional method to distinguish true or false images is to embed digital watermarks or signatures in digital images in advance. However, it still has a lot of weaknesses. Therefore it forms a new research direction to authenticate digital images that are embedded with no digital watermarks or signatures.
Probably, an altered digital image cannot be authenticated. But, more often than not, some invisible clues may be left in the tampering process, because the tampering may disturb some underlying statistical properties of images. On account of this assumption, we propose two methods to detect tampered images in terms of different tampering forms:
(1) A detection method based on the statistical properties of double-JPEG compression. This approach is used to detect previous quantification coefficients of double-JPEG compressed image pieces. First, identify some features of DCT histograms of particular coefficients due to double compression. Second, use SVM to estimate the original quantification coefficients from double-compressed images. The image areas with different quantification coefficients are worth our questioning. Finally, basing on many different sets of double-JPEG compressed images, use a series of experiments to verify the performance and reliability of this method.
(2) A detection method based on image matching. This detection method focuses on the tampering type which is copying an image area then moving it to cover other important parts (Copy-Move forgery image). According to tampered image's sizes, two checking algorithms are put forth and their effectiveness and accountability are
一幅被改动过的数字图像可能用肉眼无法辨别真伪，但常常会因为篡改而留下一些不可见的线索。篡改会干扰图像一些潜在的统计特性。基于这样的设想，针对不同的图像篡改方式，本文提出了二种检测篡改图像的方法：
(1)基于双JPEG压缩统计特性的检测方法。这种方法是通过检测双JPEG压缩图像的图像块先前的量化系数来实现的。首先，定义了在两次压缩时，出现在特定位置上的DCT系数直方图的一些特性。然后，使用支持向量机来估计双压图像的初始的量化系数，有不同量化系数的图像块便是值得怀疑的区域。最后，基于大量不同的双压图像集，用一系列的实验来验证本方法的性能及可靠性。
(2)基于图像匹配的检测方法。此检测方法着重针对复制图像本身的一部分来遮盖另一重要部分的篡改类型(复制遮盖篡改)，针对不同尺寸大小的篡改图像提出了两种检测算法，并论证了检测算法的有效性和可靠性。这些方法对于复制区域被增强或修饰使之融入到背景之中，依然非常有效。
对于每一种方法，都论述了它的理论基础，并在实验结果中，展示了它在检测篡改图像时的有效性。
As a result of the widespread use of powerful image-processing software, it is increasingly easier to tamper digital images. For the time being, it has been possible to add or remove some important people or things to or from an image, leaving behind no obvious traces at the same time. Since digital cameras and digital vidicons replaced traditional analogue counterparts, it has become more and more urgent to identify digital images, to confirm image contents and detect their tampering. The traditional method to distinguish true or false images is to embed digital watermarks or signatures in digital images in advance. However, it still has a lot of weaknesses. Therefore it forms a new research direction to authenticate digital images that are embedded with no digital watermarks or signatures.
Probably, an altered digital image cannot be authenticated. But, more often than not, some invisible clues may be left in the tampering process, because the tampering may disturb some underlying statistical properties of images. On account of this assumption, we propose two methods to detect tampered images in terms of different tampering forms:
(1) A detection method based on the statistical properties of double-JPEG compression. This approach is used to detect previous quantification coefficients of double-JPEG compressed image pieces. First, identify some features of DCT histograms of particular coefficients due to double compression. Second, use SVM to estimate the original quantification coefficients from double-compressed images. The image areas with different quantification coefficients are worth our questioning. Finally, basing on many different sets of double-JPEG compressed images, use a series of experiments to verify the performance and reliability of this method.
(2) A detection method based on image matching. This detection method focuses on the tampering type which is copying an image area then moving it to cover other important parts (Copy-Move forgery image). According to tampered image's sizes, two checking algorithms are put forth and their effectiveness and accountability are
引文
[1] H. Farid. Creating and detecting doctored and virtual images: Implications to the childpornography prevention act. Technical Report TR2004-518, Dartmouth College, September 2004.
[2] M.Wu and B. Liu.Watermarking for image authentication. In IEEE International Conference on Image Processing, volume 2, pages 437-441, 1998.
[3] G. L. Friedman. The trustworthy camera: Restoring credibility to the photographic image.IEEE Transactions on Consumer Electronics, 39(4):905-910, 1993.
[4] P. W. Wong. A public key watermark for image verification and authentication. In IEEE International Conference on Image Processing, volume 1, pages 455-459, 1998.
[5] C. W. Honsinger, P. Jones, M. Rabbani, and J. C. Stoffel. Lossless recovery of an original image containing embedded data. U.S. Patent Application, Docket No. 77102/E-D, 1999.
[6] J. Fddrich, M. Goljan, and M. Du. Invertible authentication. In Proceedings of SPIE, Security and Watermarking of Multimedia Contents, 2001.
[7] C.-Y. Lin and S.-F. Chang. A robust image authentication algorithm surviving JPEG lossy compression. In SPIE Storage and Retrieval of Image/Video Databases, volume 3312, pages 296-307, 1998.
[8] M. Schneider and S.-F. Chang. A robust content-based digital signature for image authentication. In IEEE International Conference on Image Processing, volume 2, pages 227-230,1996.
[9] D. Storck. A new approach to integrity of digital images. In IFIP Conference on Mobile Communication, pages 309-316, 1996.
[10] B. M. Macq and J.-J. Quisquater. Cryptology for digital TV broadcasting. Proceedings of the IEEE, 83(6):944-957, 1995.
[11] S. Bhattacharjee and M. Kutter. Compression-tolerant image authentication. In IEEE International Conference on Image Processing, volume 1, pages 435-439, 1998.
[12] D. Kundur and D. Hatzinakos. Digital watermarking for tell-tale tamper proofing and authentication. Proceedings of the IEEE, 87(7):1167-1180, 1999.
[13] M.Wu and B. Liu.Watermarking for image authentication. In IEEE International Conference on Image Processing, volume 2, pages 437-441, 1998.
[14] P. W. Wong. A watermark for image integrity and ownership verification. In Proceedings of IS&TPIC Conference, pages 374-379, Portland, Oregon, May 1998.
[15] J. Fridrich. Security of fragile authentication watermarks with localization. In Proceedings of SPIE, Electronic Imaging 2002, Security and Watermarking of Multimedia Contents, volume 4675,pages 691-700, January 2002.
[16] M. U. Celik, G. Sharma, E. Saber, and A. M. Tekalp. Hierarchical watermarking for secure image authentication with localization. IEEE Transactions on Image Processing, 11(6):585-595,June 2002.
[17] M. Yeung and F. Mintzer. An invisible watermarking technique for image verification. In Proceedings of the International Conference on Image Processing, volume 1, pages 680- 683,1997.
[18] J. Fridrich and M. Goljan. Images with self-correcting capabilities. In Proceedings of the??IEEE International Conference on Image Processing, volume 3, pages 792-796, 1999.
[19] E. T. Lin, C. I. Podilchuk, and E. J. Delp. Detection of image alterations using semi-fragile watermarks. In Proceedings of the SPIE International Conference on Security and Watermarking of Multimedia Contents Ⅱ, volume 3971, pages 152-163, 2000.
[20] D. Kundur and D. Hatzinakos. Blind image deconvolution. IEEE Signal Processing Magazine, 13(3):43-64, 1996.
[21] D. Kundur and D. Hatzinakos. Semi-blind image restoration based on tell-tale watermarking.In Conference Record of the 32nd Asilomar Conference on Signals, Systems, and Computers,volume 2, pages 933-937, 1998.
[22] P. Blythe and J. Fridrich. Secure digital camera. In Digital Forensic Research Workshop, Baltimore, Maryland, August 2004.
[23] S. A. Craver, M. Wu, B. Liu, A. Stubblefield, B. Swartzlander, D. S. Wallach, D. Dean, and E. W. Felten, "Reading between the lines: Lessons from the SDMI challenge,"10thUSENIX Security Symposium, Washington, D.C., Aug 13-17, 2001.
[24] S. A. Craver, M. Wu, B. Liu, A. Stubblefield, B. Swartzlander, and D. S. Wallach. Reading between the lines: Lessons from the SDMI challenge. In 10th USENIX Security Symposium,Washington DC, 2001.
[25] H. Farid. Detecting digital forgeries using bispectral analysis. Technical Report AIM-1657, Massachusetts Institue of Technology, August 1999.
[26] T.-T. Ng and S.-F. Chang. Blind detection of digital photomontage using higher order statistics.Technical Report ADVENT 201-2004-1, Columbia University, June 2004.
[27] A. C. Popescu and H. Farid. Statistical tools for digital forensics. In Proceedings of the 6th Information Hiding Workshop, May 2004.
[28] S. Lyu and H. Farid. How realistic is photorealistic? IEEE Transactions on Signal Processing,2005. (in press).
[29] J. Fridrich, D. Soukal, and J. Luk. Detection of copy-move forgery in digital images. In Proceedings of Digital Forensic Research Workshop, August 2003.
[30] Jan Luk and Jessica Fridrich, Estimation of Primary Quantization Matrix in Double Compressed JPEG Images, In Digital Forensic Research Workshop, Cleveland, Ohio, August 2003..
[31] Fan Z. and de Queiroz R. L.: "Identification of Bitmap Compression History: JPEG Detection and Quantizer Estimation", in IEEE Transactions on Image Processing, vol. 12, No. 2, February 2003, pp. 230-235.
[32] B. Scholkopf, J. C. Platt, J. T. Shawe, A. J. Smola, R. C.Williamson,"Estimation the support of a high-dimensional Distribution", Technical Report MSR-TR-99-87,Microsoft Research
[33] Yunqiang Chen, Xiang Zhou, and Thomas S. Huang,"one-class SVM for learning in image retrieval",In Proc. IEEE Int'l Conf on lmage Processing 2001, Thessaloniki, Greece
[34] 边肇祺，张学工等．模式识别[M]．清华大学出版社，2001．5
[35] Chih-Chung Chang and Chih-Jen Lin, LIBSVM : a library for support vector machines, 2001
[36] C.-W. Hsu and C.-J. Lin. A comparison of methods for multi-class support vector machines, IEEE Transactions on Neural Networks, 13(2002), 415-425.
[37] SVM and Kernel Methods Matlab Toolbox,Modified by A. Rakotomamonjy on 20/12/2005.http://asi.insa-rouen.fr/～arakotom/toolbox/index.html[38] Jessica Fridrich, David Soukal, and Jan Luk, Detection of Copy-Move Forgery in Digital Images, http://www.ws.binghamton.edu/fridrich/Research/copymove.pdf
[39] Alin C Popescu and Hany Farid, Exposing Digital Forgeries by Detecting Duplicated Image Regions, http://www.ists.dartmouth.edu/library/tr-2004-515.pdf
[40] R. Duda and P. Hart. Pattern Classification and Scene Analysis. John Wiley and Sons, 1973.
[2] M.Wu and B. Liu.Watermarking for image authentication. In IEEE International Conference on Image Processing, volume 2, pages 437-441, 1998.
[3] G. L. Friedman. The trustworthy camera: Restoring credibility to the photographic image.IEEE Transactions on Consumer Electronics, 39(4):905-910, 1993.
[4] P. W. Wong. A public key watermark for image verification and authentication. In IEEE International Conference on Image Processing, volume 1, pages 455-459, 1998.
[5] C. W. Honsinger, P. Jones, M. Rabbani, and J. C. Stoffel. Lossless recovery of an original image containing embedded data. U.S. Patent Application, Docket No. 77102/E-D, 1999.
[6] J. Fddrich, M. Goljan, and M. Du. Invertible authentication. In Proceedings of SPIE, Security and Watermarking of Multimedia Contents, 2001.
[7] C.-Y. Lin and S.-F. Chang. A robust image authentication algorithm surviving JPEG lossy compression. In SPIE Storage and Retrieval of Image/Video Databases, volume 3312, pages 296-307, 1998.
[8] M. Schneider and S.-F. Chang. A robust content-based digital signature for image authentication. In IEEE International Conference on Image Processing, volume 2, pages 227-230,1996.
[9] D. Storck. A new approach to integrity of digital images. In IFIP Conference on Mobile Communication, pages 309-316, 1996.
[10] B. M. Macq and J.-J. Quisquater. Cryptology for digital TV broadcasting. Proceedings of the IEEE, 83(6):944-957, 1995.
[11] S. Bhattacharjee and M. Kutter. Compression-tolerant image authentication. In IEEE International Conference on Image Processing, volume 1, pages 435-439, 1998.
[12] D. Kundur and D. Hatzinakos. Digital watermarking for tell-tale tamper proofing and authentication. Proceedings of the IEEE, 87(7):1167-1180, 1999.
[13] M.Wu and B. Liu.Watermarking for image authentication. In IEEE International Conference on Image Processing, volume 2, pages 437-441, 1998.
[14] P. W. Wong. A watermark for image integrity and ownership verification. In Proceedings of IS&TPIC Conference, pages 374-379, Portland, Oregon, May 1998.
[15] J. Fridrich. Security of fragile authentication watermarks with localization. In Proceedings of SPIE, Electronic Imaging 2002, Security and Watermarking of Multimedia Contents, volume 4675,pages 691-700, January 2002.
[16] M. U. Celik, G. Sharma, E. Saber, and A. M. Tekalp. Hierarchical watermarking for secure image authentication with localization. IEEE Transactions on Image Processing, 11(6):585-595,June 2002.
[17] M. Yeung and F. Mintzer. An invisible watermarking technique for image verification. In Proceedings of the International Conference on Image Processing, volume 1, pages 680- 683,1997.
[18] J. Fridrich and M. Goljan. Images with self-correcting capabilities. In Proceedings of the??IEEE International Conference on Image Processing, volume 3, pages 792-796, 1999.
[19] E. T. Lin, C. I. Podilchuk, and E. J. Delp. Detection of image alterations using semi-fragile watermarks. In Proceedings of the SPIE International Conference on Security and Watermarking of Multimedia Contents Ⅱ, volume 3971, pages 152-163, 2000.
[20] D. Kundur and D. Hatzinakos. Blind image deconvolution. IEEE Signal Processing Magazine, 13(3):43-64, 1996.
[21] D. Kundur and D. Hatzinakos. Semi-blind image restoration based on tell-tale watermarking.In Conference Record of the 32nd Asilomar Conference on Signals, Systems, and Computers,volume 2, pages 933-937, 1998.
[22] P. Blythe and J. Fridrich. Secure digital camera. In Digital Forensic Research Workshop, Baltimore, Maryland, August 2004.
[23] S. A. Craver, M. Wu, B. Liu, A. Stubblefield, B. Swartzlander, D. S. Wallach, D. Dean, and E. W. Felten, "Reading between the lines: Lessons from the SDMI challenge,"10thUSENIX Security Symposium, Washington, D.C., Aug 13-17, 2001.
[24] S. A. Craver, M. Wu, B. Liu, A. Stubblefield, B. Swartzlander, and D. S. Wallach. Reading between the lines: Lessons from the SDMI challenge. In 10th USENIX Security Symposium,Washington DC, 2001.
[25] H. Farid. Detecting digital forgeries using bispectral analysis. Technical Report AIM-1657, Massachusetts Institue of Technology, August 1999.
[26] T.-T. Ng and S.-F. Chang. Blind detection of digital photomontage using higher order statistics.Technical Report ADVENT 201-2004-1, Columbia University, June 2004.
[27] A. C. Popescu and H. Farid. Statistical tools for digital forensics. In Proceedings of the 6th Information Hiding Workshop, May 2004.
[28] S. Lyu and H. Farid. How realistic is photorealistic? IEEE Transactions on Signal Processing,2005. (in press).
[29] J. Fridrich, D. Soukal, and J. Luk. Detection of copy-move forgery in digital images. In Proceedings of Digital Forensic Research Workshop, August 2003.
[30] Jan Luk and Jessica Fridrich, Estimation of Primary Quantization Matrix in Double Compressed JPEG Images, In Digital Forensic Research Workshop, Cleveland, Ohio, August 2003..
[31] Fan Z. and de Queiroz R. L.: "Identification of Bitmap Compression History: JPEG Detection and Quantizer Estimation", in IEEE Transactions on Image Processing, vol. 12, No. 2, February 2003, pp. 230-235.
[32] B. Scholkopf, J. C. Platt, J. T. Shawe, A. J. Smola, R. C.Williamson,"Estimation the support of a high-dimensional Distribution", Technical Report MSR-TR-99-87,Microsoft Research
[33] Yunqiang Chen, Xiang Zhou, and Thomas S. Huang,"one-class SVM for learning in image retrieval",In Proc. IEEE Int'l Conf on lmage Processing 2001, Thessaloniki, Greece
[34] 边肇祺，张学工等．模式识别[M]．清华大学出版社，2001．5
[35] Chih-Chung Chang and Chih-Jen Lin, LIBSVM : a library for support vector machines, 2001
[36] C.-W. Hsu and C.-J. Lin. A comparison of methods for multi-class support vector machines, IEEE Transactions on Neural Networks, 13(2002), 415-425.
[37] SVM and Kernel Methods Matlab Toolbox,Modified by A. Rakotomamonjy on 20/12/2005.http://asi.insa-rouen.fr/～arakotom/toolbox/index.html[38] Jessica Fridrich, David Soukal, and Jan Luk, Detection of Copy-Move Forgery in Digital Images, http://www.ws.binghamton.edu/fridrich/Research/copymove.pdf
[39] Alin C Popescu and Hany Farid, Exposing Digital Forgeries by Detecting Duplicated Image Regions, http://www.ists.dartmouth.edu/library/tr-2004-515.pdf
[40] R. Duda and P. Hart. Pattern Classification and Scene Analysis. John Wiley and Sons, 1973.