آشکارسازی فشرده‌سازی JPEG مضاعف با استفاده از شبکه‌های عصبی عمیق در حوزه مکان

محورهای موضوعی : انرژی های تجدیدپذیر

محمد رحمتی ¹ , فربد رزازی ² , علیرضا بهراد ³

1 - دانشکده مهندسی برق و کامپیوتر- واحد علوم و تحقیقات، دانشگاه آزاد اسلامی، تهران، ایران
2 - دانشکده مهندسی برق و کامپیوتر- واحد علوم و تحقیقات، دانشگاه آزاد اسلامی، تهران، ایران
3 - دانشکده فنی و مهندسی - دانشگاه شاهد، تهران، ایران

تاریخ دریافت : 1400/02/20 تاریخ پذیرش : 1400/05/22 تاریخ انتشار : 1400/06/01

کلید واژه: فیلتر تطبیقی, صحت آشکارسازی, شبکه عصبی پیچشی, خودرمزگذار پیچشی, مکان‌یابی محل دست‌کاری,

چکیده مقاله :

: با افزایش علاقه‌مندی به فشرده نمودن تصاویر با فرمت فرمت گروه مشترک متخصصان عکاسی (JPEG)، یکی از مهم‌ترین مباحث در دست‌کاری تصاویر دیجیتال، یافتن روشی مناسب جهت آشکارسازی فشرده‌سازی JPEG مضاعف است. در این مقاله با معرفی یک فیلتر تطبیقی آموزش‌دیده بر پایه خودرمزگذار پیچشی (CAE) و در حوزه مکان، به این موضوع پرداخته می‌شود تا با حذف اطلاعات تداخلی ناشی از محتوای تصویر، آشکارسازی دقیق‌تری داشته باشیم. از آنجایی که شبکه عصبی پیچشی (CNN) توانسته عملکرد موفقی در طبقه‌بندی تصاویر داشته باشد، از این شبکه‌ها در قسمت طبقه‌بندی استفاده می‌شود. مدل پیشنهادی بر اساس CAE متوالی شده با CNN است که توانسته دقت آشکارسازی و حساسیت به ضرایب کیفیت (QFs) قابل قبولی را در دو سناریوی هم‌تراز و ناهم‌تراز ارائه نماید. این مدل توانسته در برخی از حالت ها، حساسیت نسبت به ضرایب کیفیت را تا 86 در صد در مقدار کاهش خطای نسبی (RER) بهبود دهد. آزمایش‌های دیگری از جمله مکان‌یابی محل دست‌کاری بر روی مجموعه داده RAISE برای ارزیابی روش پیشنهادی انجام شده است. این نتایج نشان‌دهنده عملکرد بسیار خوب این روش نسبت به الگوریتم‌های مشابه در شرایطی است که ضریب کیفیت فشرده‌سازی دوم بزرگ‌تر از ضریب کیفیت فشرده‌سازی اول باشد.

چکیده انگلیسی:

With the increasing interest in Joint Photographic Experts Group (JPEG) image compression, one of the most important issues in digital image manipulation is finding a proper method to detect double JPEG compression. This paper introduces a trained adaptive filter based on spatial-domain convolutional autoencoder (CAE). This filter can remove interference information caused by image content to have a more accurate detection. The convolutional neural network (CNN) has been widely employed for accurate image classification; therefore, a CNN is used in the classification part of the proposed algorithm. The proposed model is based on consecutive CAE with CNN, which is able to provide acceptable detection accuracy and sensitivity to quality factors (QFs) in two scenarios, i.e. aligned and non-aligned forgeries. This model improves the sensitivity to quality factors by up to 86% in the relative error reduction (RER) rate in some cases. Other experiments such as manipulation localization on the RAISE dataset have been performed to evaluate the proposed method. These results show the superior performance of this method compared to similar algorithms in the situations that the quality factor of the second compression is greater the quality factor of the first compression. 

منابع و مأخذ:

[1] A. Faulkner, C. Chavez, “Adobe photoshop CC classroom in a book (2018 release)”, 1^th Edition, Adobe Press, 2017.

[2] W. Zhang, “Smartphone photography in urban china”, International Journal of Social, Behavioral, Educational, Economic, Business and Industrial Engineering, vol. 11, no. 1, pp. 231–239, 2017 (doi: 10.5281/zenodo.1128985).

[3] V. Verma, D. Singh, N. Khanna, “Block-level double JPEG compression detection for image forgery localization”, arXiv preprint arXiv:2003.09393, 2020.

[4] W. Ahn, S.H. Nam, M. Son, H.K. Lee, S. Choi, “End-to-end double JPEG detection with a 3D convolutional network in the DCT domain”, Electronics Letters, vol. 56, no. 2, pp. 82-85, 2020 (doi: 10.1049/el.2019.2719).

[5] C. Deng, Z. Li, X. Gao, D. Tao, “Deep multi-scale discriminative networks for double jpeg compression forensics”, ACM Transactions on Intelligent Systems and Technology, vol. 10, no. 2, pp. 1-20, 2019 (doi: 10.1­1­45/3­301274).

[6] M. Barni, L. Bondi, N. Bonettini, P. Bestagini, A. Costanzo, M. Maggini, B. Tondi, S. Tubaro, “Aligned and non-aligned double JPEG detection using convolutional neural networks”, Journal of Visual Communication and Image Representation, vol. 49, pp.153-163, 2017 (doi: 10.1016/j.jvcir.2017.09.003).

[7] P. Korus, J. Huang, “Multi-scale fusion for improved localization of malicious tampering in digital images”,  IEEE Trans. on Image Processing, vol. 25, no. 3, pp. 1312-1326, 2016 (doi: 10.1109/TIP.2016.2518870).

[8] C. Pasquini, P. Schöttle, R. Böhme, G. Boato, F. Pèrez-Gonzàlez, “Forensics of high quality and nearly identical jpeg image recompression”, Proceedings of theIH&MMSec, pp. 11-21, Spain,June 2016 (doi: 10.1145/2909827.2930787).

[9] B. Li, H. Zhang, H. Luo, S. Tan, “Detecting double JPEG compression and its related anti-forensic operations with CNN”, Multimedia Tools and Applications, vol. 78, no. 7, pp. 8577-8601, 2019 (doi: 10.1007/s11042-018-7073-3).

[10] L. Bondi, D. Güera, L. Baroffio, P. Bestagini, E.J. Delp, S. Tubaro, “A preliminary study on convolutional neural networks for camera model identification”, Electronic Imaging, no. 7, pp.67-76, 2017 (doi: 10.2352/I­SS­N.2470-1173.2017.7.MWSF-327).

[11] B. Bayar, M.C. Stamm, “A deep learning approach to universal image manipulation detection using a new convolutional layer”, Proceedings of the IH&MMSec, pp. 5-10, Spain, June 2016 (doi: 10.1145/2909827.­2­9­3­­0786).

[12] Q. Wang, R. Zhang, “Double JPEG compression forensics based on a convolutional neural network”, EURASIP Journal on Information Security, Article number: 23, pp. 1-12, Oct. 2016 (doi: 10.1186/s13635-016-0047-y).

[13] I. Amerini, L. Ballan, R. Caldelli, A.D. Bimbo, G. Serra, “A sift-based forensic method for copy–move attack detection and transformation recovery”, IEEE Trans. on Information Forensics and Security, vol. 6, no. 3, pp. 1099-1110, Sept. 2011 (doi: 10.1109/TIFS.2011.2129512).

[14] D.G. Lowe, “Distinctive image features from scale-invariant keypoints”, International Journal of Computer Vision, vol. 60, no. 2, pp. 91-110, 2004 (doi: 10.1023/B:VISI.0000029664.99615.94).

[15] A. Taimori, F. Razzazi, A. Behrad, A. Ahmadi, M. Babaie-Zadeh, “Quantization-unaware double JPEG compression detection”, Journal of Mathematical Imaging and Vision, vol. 54, no. 3, pp. 269-286, 2016 (doi: 10.1007/s10851-015-0602-z).

[16] T. Bianchi, A. Piva, “Image forgery localization via block-grained analysis of JPEG artifacts”, IEEE Trans. on Information Forensics and Security, vol. 7, no. 3, pp. 1003-1017, 2012 (doi: 10.1109/TIFS.2012.2187516).

[17] T. Mahmood, M. Shah, J. Rashid, T. Saba, M. W. Nisar, M. Asif, “A passive technique for detecting copy-move forgeries by image feature matching”, Multimedia Tools and Applications, vol. 79, no. 43, pp. 31759-31782, Aug. 2020 (doi: 10.1007/s11042-020-09655-2).

[18] J. Wang, W. Huang, X. Luo, Y.Q. Shi, S.K. Jha, “Non-aligned double JPEG compression detection based on refined Markov features in QDCT domain”, Journal of Real-Time Image Processing, vol. 17, no. 1, pp. 7-16, 2020 (doi: 10.1007/s11554-019-00929-z).

[19] M.K. Mihcak, I. Kozintsev, K. Ramchandran, “Spatially adaptive statistical modeling of wavelet image coefficients and its application to denoising”, Proceeding of the IEEE/ICASS, vol. 6, pp. 3253-3256, Phoenix, AZ, USA, March 1999 (doi: 10.1109/ICASSP.1999.757535).

[20] V. Verma, N. Agarwal, N. Khanna, “DCT-domain deep convolutional neural networks for multiple JPEG compression classification”, Signal Processing: Image Communication, vol. 67, pp. 22-33, Sept. 2018 (doi: 10.1016­/j­.i­mage.2018.04.014).

[21] B. Li, H. Zhang, H. Luo, S. Tan, “Detecting double JPEG compression and its related anti-forensic operations with CNN”, Multimedia Tools and Applications, vol 78, no. 7, pp. 8577-8601, Sept. 2019 (doi: 10.1007/s11­04­2-0­18-7073-3).

[22] Y. Rao, J. Ni, “A deep learning approach to detection of splicing and copy-move forgeries in images”, Proceeding of the IEEE/WIFS, pp. 1-6, Abu Dhabi, United Arab Emirates, Dec. 2016 (doi: 10.1109/WIF­S.2016.­7823911).

[23] J. Fridrich, J. Kodovsky, "Rich models for steganalysis of digital images", IEEE Trans. on Information Forensics and Security, vol. 7, no. 3, pp. 868-882, June 2012 (doi: 10.1109/TIFS.2012.2190402).

[24] J. Chen, X. Kang, Y. Liu, Z.J. Wang, “Median filtering forensics based on convolutional neural networks”, IEEE Signal Processing Letters, vol. 22. No. 11, pp.1849-1853, 2015 (doi: 10.1109/LSP.2015.2438008).

[25] B. Bayar, M.C. Stamm, “Constrained convolutional neural networks: a new approach towards general purpose image manipulation detection”, IEEE Trans. on Information Forensics and Security, vol. 13, no. 11, pp.2691-2706, 2018 (doi: 10.1109/TIFS.2018.2825953).

[26] L. Gondara, “Medical image denoising using convolutional denoising autoencoders”, Proceeding of the IEEE/ICDMW, pp. 241-246, Barcelona, Spain, Dec. 2016 (doi: 10.1109/ICDMW.2016.0041).

[27] K. He, X. Zhang, S. Ren, J. Sun, “Deep residual learning for image recognition”, Proceedings of the IEEE/CVPR, pp. 770-778, Las Vegas, USA, June 2016 (doi: 10.1109/CVPR.2016.90).

[28] T. Ridnik, H. Lawen, A. Noy, I. Friedman, “TResNet: High performance GPU-dedicated architecture”, arXiv preprint arXiv:2003.13630, 2020.

[29] Y. Yousfi, J.B.E. Khvedchenya, J. Fridrich, “ImageNet pretrained CNNs for JPEG steganalysis”, Proceeding of the IEEE/WIFS, pp. 1-6, New York, USA, Dec. 2020 (doi: 10.1109/WIFS49906.2020.9360897).

[30] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, A. Rabinovich, “Going deeper with convolutions”, Proceedings of the IEEE/CVPR, pp. 1-9, Boston, USA, June 2015 (doi: 10.1109/CVPR.2015.7298594).

[31] D.A. Clevert, T. Unterthiner, S. Hochreiter, “Fast and accurate deep network learning by exponential linear units (elus)”, arXiv preprint arXiv:1511.07289, 2015.

[32] S. Ioffe, C. Szegedy, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” arXiv preprint arXiv:1502.03167, 2015.

[33] D.T. Dang-Nguyen, C. Pasquini, V. Conotter, G. Boato, “RAISE: A raw images dataset for digital image forensics”, Proceedings of the MMSys, pp. 219-224, March 2015 (doi: 10.1145/2713168.2713194).

[34] K. He, X. Zhang, S. Ren, J. Sun, “Delving deep into rectifiers: surpassing human-level performance on imagenet classification”, Proceedings of the IEEE/CV, pp. 1026-1034, Santiago, Chile, Dec.2015 (doi: 10.1109/ICCV.2015.123).

[35] D.P. Kingma, J. Ba, “Adam: A method for stochastic optimization”, arXiv preprint arXiv:1412.6980, 2014.

[36] H. Yao, H. Wei, C. Qin, X. Zhang, “An improved first quantization matrix estimation for nonaligned double compressed JPEG images”,  Signal Processing, vol. 170, Article Number: 107430, May 2020 (doi: 10.1016/­j.sig­pro.2019.107430).