首页 » 文章 » 文章详细信息
Advances in Multimedia Volume 2017 ,2017-01-12
Revealing Traces of Image Resampling and Resampling Antiforensics
Research Article
Anjie Peng 1 , 2 Yadong Wu 1 Xiangui Kang 2
Show affiliations
DOI:10.1155/2017/7130491
Received 2016-08-19, accepted for publication 2016-12-12, Published 2016-12-12
PDF
摘要

Image resampling is a common manipulation in image processing. The forensics of resampling plays an important role in image tampering detection, steganography, and steganalysis. In this paper, we proposed an effective and secure detector, which can simultaneously detect resampling and its forged resampling which is attacked by antiforensic schemes. We find that the interpolation operation used in the resampling and forged resampling makes these two kinds of image show different statistical behaviors from the unaltered images, especially in the high frequency domain. To reveal the traces left by the interpolation, we first apply multidirectional high-pass filters on an image and the residual to create multidirectional differences. Then, the difference is fit into an autoregressive (AR) model. Finally, the AR coefficients and normalized histograms of the difference are extracted as the feature. We assemble the feature extracted from each difference image to construct the comprehensive feature and feed it into support vector machines (SVM) to detect resampling and forged resampling. Experiments on a large image database show that the proposed detector is effective and secure. Compared with the state-of-the-art works, the proposed detector achieved significant improvements in the detection of downsampling or resampling under JPEG compression.

授权许可

Copyright © 2017 Anjie Peng et al. 2017
This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

图表

Example of the upscaling (s=3/2, bilinear) process for the ith row of image B (the first line). The corresponding interpolated weights are shown in the bracket.

Top row: unaltered image, p-map, and its Fourier spectrum. Second row: upscaled image (s=3/2, bilinear). Third row: forged upscaled image (attack 1, s=3/2, bilinear, σ=0.4).

The average histogram for 1500 unaltered images, their corresponding upscaled images (s=3/2, bicubic), and forged upscaled images (attack 1, s=3/2, σ=0.4, bicubic).

The high frequency DCT subband in red shaded region is used to create R1(x,y). The coordinate (0,0) is the DC coefficient.

The distribution of FAR10 (a) and FAR11 (b) for 1500 uncompressed unaltered images and their upscaled (s=3/2, bicubic) and forged upscaled (attack 1, s=3/2, σ=0.4, bicubic) versions. X-coordinate: the index of AR coefficient; Y-coordinate: averaged value of AR coefficient.

The distribution of FAR10 (a) and FAR11 (b) for 1500 uncompressed unaltered images and their upscaled (s=3/2, bicubic) and forged upscaled (attack 1, s=3/2, σ=0.4, bicubic) versions. X-coordinate: the index of AR coefficient; Y-coordinate: averaged value of AR coefficient.

ROC curves showing detections of (a) upscaling, (b) downscaling, and (c) rotation under uncompressed scenario.

ROC curves showing detections of (a) upscaling, (b) downscaling, and (c) rotation under uncompressed scenario.

ROC curves showing detections of (a) upscaling, (b) downscaling, and (c) rotation under uncompressed scenario.

ROC curves showing detections of (a) forged upscaling, (b) forged downscaling, and (c) forged rotation under uncompressed scenario.

ROC curves showing detections of (a) forged upscaling, (b) forged downscaling, and (c) forged rotation under uncompressed scenario.

ROC curves showing detections of (a) forged upscaling, (b) forged downscaling, and (c) forged rotation under uncompressed scenario.

The 2D feature of FD (after LDA) estimated from 1500 uncompressed images of BOSSRAW database.

ROC curves of detecting resampled and forged resampled image (“ALL”) from unaltered image under uncompressed scenario.

An example showing (a) an unaltered image and tampering detection results for (b) uncompressed and (c) JPEG compressed (QF = 95) tampering. The red box of size 64 × 64 indicates that this box is predicted as tampered by the proposed detector.

An example showing (a) an unaltered image and tampering detection results for (b) uncompressed and (c) JPEG compressed (QF = 95) tampering. The red box of size 64 × 64 indicates that this box is predicted as tampered by the proposed detector.

An example showing (a) an unaltered image and tampering detection results for (b) uncompressed and (c) JPEG compressed (QF = 95) tampering. The red box of size 64 × 64 indicates that this box is predicted as tampered by the proposed detector.

通讯作者

Xiangui Kang.Guangdong Key Lab of Information Security, School of Data and Computer Science, Sun Yat-Sen University, Guangzhou, China, sysu.edu.cn.isskxg@mail.sysu.edu.cn

推荐引用方式

Anjie Peng,Yadong Wu,Xiangui Kang. Revealing Traces of Image Resampling and Resampling Antiforensics. Advances in Multimedia ,Vol.2017(2017)

您觉得这篇文章对您有帮助吗?
分享和收藏
0

是否收藏?

参考文献
[1] T. Pevný, P. Bas, J. Fridrich. (2010). Steganalysis by subtractive pixel adjacency matrix. IEEE Transactions on Information Forensics and Security.5(2):215-224. DOI: 10.1109/tsp.2004.839932.
[2] D. Vázquez-Padín, P. Comesaña, F. Pérez-González. Set-membership identification of resampled signals. :150-155. DOI: 10.1109/tsp.2004.839932.
[3] D. Vázquez-Padín, P. Comesaña. ML estimation of the resampling factor. :205-210. DOI: 10.1109/tsp.2004.839932.
[4] D. Vázquez-Padín, C. Mosquera, F. Pérez-González. Prefilter design for forensic resampling estimation. :1-6. DOI: 10.1109/tsp.2004.839932.
[5] A. C. Gallagher. Detection of linear and cubic interpolation in JPEG compressed images. :65-72. DOI: 10.1109/tsp.2004.839932.
[6] M. Kirchner, R. Böhme. (2008). Hiding traces of resampling in digital images. IEEE Transactions on Information Forensics and Security.3(4):582-592. DOI: 10.1109/tsp.2004.839932.
[7] T. Gloe, R. Böhme. Dresden image database for benchmarking digital image forensics. :22-26. DOI: 10.1109/tsp.2004.839932.
[8] H. D. Li, W. Q. Luo, X. Q. Qiu, J. W. Huang. et al.(2016). Identification of image operations based on steganalytic features. IEEE Transactions on Circuits and Systems for Video Technology. DOI: 10.1109/tsp.2004.839932.
[9] G. Cao, Y. Zhao, R. Ni. (2012). Forensic identification of resampling operators: a semi non-intrusive approach. Forensic Science International.216(1–3):29-36. DOI: 10.1109/tsp.2004.839932.
[10] S. Kay. (1988). Modern Spectral Estimation. DOI: 10.1109/tsp.2004.839932.
[11] M. Kirchner. Fast and reliable rescaling detection by spectral analysis of fixed linear predictor residue. :11-20. DOI: 10.1109/tsp.2004.839932.
[12] B. Mahdian, S. Saic. (2008). Blind authentication using periodic properties of interpolation. IEEE Transactions on Information Forensics and Security.3(3):529-538. DOI: 10.1109/tsp.2004.839932.
[13] X. Y. Feng, I. J. Cox, D. Gwenaël. (2012). Normalized energy density-based forensic detection of re-sampled images. IEEE Transactions on Multimedia.14(3):535-546. DOI: 10.1109/tsp.2004.839932.
[14] M. Kirchner, T. Gloe. On rescaling detection in re-compressed images. :21-25. DOI: 10.1109/tsp.2004.839932.
[15] H. T. Sencar, N. Memon. (2013). Digital image forensics. Counter-Forensics: Attacking Image Forensics:327-366. DOI: 10.1109/tsp.2004.839932.
[16] W. Wei, S. Wang, X. Zhang, Z. Tang. et al.(2010). Estimation of image rotation angle using interpolation-related spectral signatures with application to blind detection of image forgery. IEEE Transactions on Information Forensics and Security.5(3):507-517. DOI: 10.1109/tsp.2004.839932.
[17] J. Fridrich, J. Kodovský. (2012). Rich models for steganalysis of digital images. IEEE Transactions on Information Forensics and Security.7(3):868-882. DOI: 10.1109/tsp.2004.839932.
[18] P. Bas, T. Filler, T. Pevný. Break our steganographic system—the ins and outs of organizing BOSS. :59-70. DOI: 10.1109/tsp.2004.839932.
[19] L. Li, J. Xue, Z. Tian, N. Zheng. et al.Moment feature based forensic detection of resampled digital images. :569-572. DOI: 10.1109/tsp.2004.839932.
[20] J. Kodovský, J. Fridrich. (2014). Effect of image downsampling on steganographic security. IEEE Transactions on Information Forensics and Security.9(5):752-762. DOI: 10.1109/tsp.2004.839932.
[21] A. Peng, H. Zeng, X. Lin, X. Kang. et al.Countering anti-forensics of image resampling. :3595-3599. DOI: 10.1109/tsp.2004.839932.
[22] X. Hou, T. Zhang, G. Xiong, Y. Zhang. et al.(2014). Image resampling detection based on texture classification. Multimedia Tools and Applications.72(2):1681-1708. DOI: 10.1109/tsp.2004.839932.
[23] X. Q. Qiu, H. D. Li, W. Q. Luo, J. W. Huang. et al.A universal image forensic strategy based on steganalytic model. :165-170. DOI: 10.1109/tsp.2004.839932.
[24] D. Vázquez-Padín, C. Mosquera, F. Pérez-González. Two-dimensional statistical test for the presence of almost cyclostationarity on images. :1745-1748. DOI: 10.1109/tsp.2004.839932.
[25] A. C. Popescu, H. Farid. (2005). Exposing digital forgeries by detecting traces of resampling. IEEE Transactions on Signal Processing.53(2):758-767. DOI: 10.1109/tsp.2004.839932.
[26] C.-C. Chang, C.-J. Lin. (2011). LIBSVM: a Library for support vector machines. ACM Transactions on Intelligent Systems and Technology.2(3, article 27). DOI: 10.1109/tsp.2004.839932.
文献评价指标
浏览 108次
下载全文 32次
评分次数 0次
用户评分 0.0分
分享 0次