Image copy-move forgery detection based on multi-scale feature extraction and fusion

doi:10.11772/j.issn.1001-9081.2022081288

Abstract

Abstract:

In the field of image copy-move forgery detection， it is very challenging to locate the boundaries of tampered small objects accurately. Current deep learning-based methods locate forged regions by detecting similar content in images. However， these methods usually just transmit the final features extracted by the encoder to the decoder to generate the mask， and ignore more spatial information of forged regions contained in the high-resolution encoding features， resulting in inaccurate model output prediction results for the boundary identification of small objects. To address this problem， a detection network based on multi-scale feature extraction and fusion called SimiNet was proposed. Firstly， abundant features were extracted by the multi-scale feature extraction module. Secondly， a skip connection was added between the feature extraction module and the decoding module to bridge the gap between the encoding and decoding features， so as to identify the boundaries of small objects accurately. Finally， Log-Cosh Dice Loss function was used to take the place of cross entropy loss to reduce the impact of class-imbalance problem on detection results. Experimental results show that the F1 score of SimiNet on USCISI dataset reaches 72.54%， which is 3.39 percentage points higher than that of the suboptimal method CMSDNet （Copy-Move Similarity Detection Network）. It can be seen that SimiNet is more accurate for boundary identification of small objects and has better visualization.

Key words: copy-move, forgery detection, deep learning, skip connection, similarity detection

摘要：

在图像复制-粘贴伪造检测领域，精确地定位被篡改的小目标的边界充满挑战性。当前基于深度学习的方法通过检测图像中的相似内容来定位伪造区域，然而它们通常只是把编码器最终提取的特征传递给解码器来生成掩码，忽略了高分辨率的编码特征所包含的更多的伪造区域的空间信息，这会导致模型输出的预测结果对于小目标的边界识别不精确。针对这个问题，提出一种基于多尺度特征提取与融合的检测网络SimiNet。首先，使用多尺度特征提取模块提取丰富的特征；其次，在特征提取模块与解码模块之间添加跳跃连接以弥补编码特征与解码特征之间的差异，从而精确地识别小目标的边界；最后，用Log-Cosh Dice Loss函数替代交叉熵损失，以降低类别不平衡问题对检测结果的影响。实验结果表明，SimiNet在USCISI数据集上的F1分数达到72.54%，比次优方法CMSDNet（Copy-Move Similarity Detection Network）提升了3.39个百分点。可见，SimiNet对小目标的边界识别更精确，可视化效果更好。

关键词: 复制粘贴, 伪造检测, 深度学习, 跳跃连接, 相似性检测

CLC Number:

TP391.41

Juntao CHEN, Ziqi ZHU. Image copy-move forgery detection based on multi-scale feature extraction and fusion[J]. Journal of Computer Applications, 2023, 43(9): 2919-2924.

陈俊韬, 朱子奇. 基于多尺度特征提取与融合的图像复制-粘贴伪造检测[J]. 《计算机应用》唯一官方网站, 2023, 43(9): 2919-2924.

Figures/Tables 8

References 28

1	SABER A H， KHAN M A， MEJBEL B G. A survey on image forgery detection using different forensic approaches［J］. Advances in Science， Technology and Engineering Systems Journal， 2020， 5（3）： 361-370. 10.25046/aj050347
2	ZEDAN I A， SOLIMAN M M， ELSAYED K M， et al. Copy move forgery detection techniques： a comprehensive survey of challenges and future directions［J］. International Journal of Advanced Computer Science and Applications， 2021， 12（7）： 248-264. 10.14569/ijacsa.2021.0120729
3	KE Y， SUKTHANKAR R， HUSTON L. An efficient parts-based near-duplicate and sub-image retrieval system［C］// Proceedings of the 12th Annual ACM International Conference on Multimedia. New York： ACM， 2004： 869-876. 10.1145/1027527.1027729
4	颜普，苏亮亮，邵慧，等. 基于多支持区域局部亮度序的图像伪造检测［J］. 计算机应用， 2019， 39（9）： 2707-2711. 10.11772/j.issn.1001-9081.2019020306
	YAN P， SU L L， SHAO H， et al. Image forgery detection based on local intensity order and multi-support region［J］. Journal of Computer Applications， 2019， 39（9）： 2707-2711. 10.11772/j.issn.1001-9081.2019020306
5	林晶，黄添强，赖玥聪，等. 采用量化离散余弦变换系数检测视频单帧连续多次复制粘贴篡改［J］. 计算机应用， 2016， 36（5）： 1356-1361. 10.11772/j.issn.1001-9081.2016.05.1356
	LIN J， HUANG T Q， LAI Y C， et al. Detection of continuously and repeated copy-move forgery to single frame in videos by quantized DCT coefficients［J］. Journal of Computer Applications， 2016， 36（5）： 1356-1361. 10.11772/j.issn.1001-9081.2016.05.1356
6	ZIMBA M， SUN X M. DWT-PCA （EVD） based copy-move image forgery detection［J］. International Journal of Digital Content Technology and its Applications， 2011， 5（1）： 251-258. 10.4156/jdcta.vol5.issue1.27
7	KANG X B， WEI S M. Identifying tampered regions using singular value decomposition in digital image forensics［C］// Proceedings of the 2008 International Conference on Computer Science and Software Engineering. Piscataway： IEEE， 2008： 926-930. 10.1109/csse.2008.876
8	LEE J C， CHANG C P， CHEN W K. Detection of copy-move image forgery using histogram of orientated gradients［J］. Information Sciences， 2015， 321： 250-262. 10.1016/j.ins.2015.03.009
9	MAHMOOD T， IRTAZA A， MEHMOOD Z， et al. Copy-move forgery detection through stationary wavelets and local binary pattern variance for forensic analysis in digital images［J］. Forensic Science International， 2017， 279： 8-21. 10.1016/j.forsciint.2017.07.037
10	OUYANG J L， LIU Y Z， LIAO M. Robust copy-move forgery detection method using pyramid model and Zernike moments［J］. Multimedia Tools and Applications， 2019， 78（8）： 10207-10225. 10.1007/s11042-018-6605-1
11	YANG B， SUN X M， GUO H L， et al. A copy-move forgery detection method based on CMFD-SIFT［J］. Multimedia Tools and Applications， 2018， 77（1）： 837-855. 10.1007/s11042-016-4289-y
12	MANU V T， MEHTRE B M. Detection of copy-move forgery in images using segmentation and SURF［M］// THAMPI S M， BANDYOPADHYAY S， KRISHNAN S， et al. Advances in Signal Processing and Intelligent Recognition Systems： Proceedings of Second International Symposium on Signal Processing and Intelligent Recognition Systems， AISC 425. Cham： Springer， 2016： 645-654. 10.1007/978-3-319-28658-7_55
13	焦丽鑫，杜振龙. 基于均值漂移的图像复制粘贴伪造盲检测［J］. 计算机应用， 2014， 34（3）： 806-809.
	JIAO L X， DU Z L. Copy-paste image forgery blind detection based on mean shift［J］. Journal of Computer Applications， 2014， 34（3）： 806-809.
14	MEENA K B， TYAGI V. A hybrid copy-move image forgery detection technique based on Fourier-Mellin and scale invariant feature transforms［J］. Multimedia Tools and Applications， 2020， 79（11/12）： 8197-8212. 10.1007/s11042-019-08343-0
15	RAO Y， NI J Q. A deep learning approach to detection of splicing and copy-move forgeries in images［C］// Proceedings of the 2016 IEEE International Workshop on Information Forensics and Security. Piscataway： IEEE， 2016： 1-6. 10.1109/wifs.2016.7823911
16	OUYANG J L， LIU Y Z， LIAO M. Copy-move forgery detection based on deep learning［C］// Proceedings of the 10th International Congress on Image and Signal Processing， BioMedical Engineering and Informatics. Piscataway： IEEE， 2017： 1-5. 10.1109/cisp-bmei.2017.8301940
17	WU Y， ABD-ALMAGEED W， NATARAJAN P. Image copy-move forgery detection via an end-to-end deep neural network［C］// Proceedings of the 2018 IEEE Winter Conference on Applications of Computer Vision. Piscataway： IEEE， 2018： 1907-1915. 10.1109/wacv.2018.00211
18	WU Y， ABD-ALMAGEED W， NATARAJAN P. BusterNet： detecting copy-move image forgery with source/target localization［C］// Proceedings of the 2018 European Conference on Computer Vision， LNCS 11210. Cham： Springer， 2018： 170-186.
19	CHEN B J， TAN W J， COATRIEUX G， et al. A serial image copy-move forgery localization scheme with source/target distinguishment［J］. IEEE Transactions on Multimedia， 2021， 23： 3506-3517. 10.1109/tmm.2020.3026868
20	SIMONYAN K， ZISSERMAN A. Very deep convolutional networks for large-scale image recognition［EB/OL］. （2015-04-10）［2022-07-20］..
21	SZEGEDY C， LIU W， JIA Y Q， et al. Going deeper with convolutions［C］// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2015： 1-9. 10.1109/cvpr.2015.7298594
22	CHEN L C， PAPANDREOU G， SCHROFF F， et al. Rethinking atrous convolution for semantic image segmentation［EB/OL］. （2017-12-05）［2022-07-20］.. 10.1007/978-3-030-01234-2_49
23	RONNEBERGER O， FISCHER P， BROX T. U-net： convolutional networks for biomedical image segmentation［C］// Proceedings of the 2015 International Conference on Medical Image Computing and Computer-Assisted Intervention， LNCS 9351. Cham： Springer， 2015： 234-241.
24	JADON S. A survey of loss functions for semantic segmentation［C］// Proceedings of the 2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology. Piscataway： IEEE， 2020： 1-7. 10.1109/cibcb48159.2020.9277638
25	DONG J， WANG W， TAN T N. CASIA image tampering detection evaluation database［C］// Proceedings of the 2013 IEEE China Summit and International Conference on Signal and Information Processing. Piscataway： IEEE， 2013： 422-426. 10.1109/chinasip.2013.6625374
26	TRALIC D， ZUPANCIC I， GRGIC S， et al. CoMoFoD - new database for copy-move forgery detection［C］// Proceedings of the ELMAR-2013. Piscataway： IEEE， 2013： 49-54.
27	COZZOLINO D， POGGI G， VERDOLIVA L. Efficient dense-field copy-move forgery detection［J］. IEEE Transactions on Information Forensics and Security， 2015， 10（11）： 2284-2297. 10.1109/tifs.2015.2455334
28	LI Y M， ZHOU J T. Fast and effective image copy-move forgery detection via hierarchical feature point matching［J］. IEEE Transactions on Information Forensics and Security， 2019， 14（5）： 1307-1322. 10.1109/tifs.2018.2876837

数据集	样本数	图像分辨率
USCISI	100 000	340×260~1 068×944
CASIA	1 313	240×160~900×600
CoMoFoD	5 000	512×512

数据集	样本数	图像分辨率
USCISI	100 000	340×260~1 068×944
CASIA	1 313	240×160~900×600
CoMoFoD	5 000	512×512

模型	模型组成	F1/%
模型1	base	59.70
模型2	base+skip	63.89
模型3	base+multi-channel	69.24
模型4	base+skip+multi-channel	70.40

模型	模型组成	F1/%
模型1	base	59.70
模型2	base+skip	63.89
模型3	base+multi-channel	69.24
模型4	base+skip+multi-channel	70.40

损失函数	BusterNet-simi	CMSDNet	SimiNet
交叉熵	52.36	69.15	70.40
Log-Cosh Dice Loss	55.92	71.23	72.54