基于图像噪声残差的数字图像来源取证

doi:10.16180/j.cnki.issn1007-7820.2020.02.007

电子科技 ›› 2020, Vol. 33 ›› Issue (2): 37-42.doi: 10.16180/j.cnki.issn1007-7820.2020.02.007

基于图像噪声残差的数字图像来源取证

黄明瑛

杭州电子科技大学计算机学院,浙江杭州 310018

收稿日期:2019-01-15 出版日期:2020-02-15 发布日期:2020-03-12
作者简介:黄明瑛(1991-),男,硕士研究生。研究方向:图像识别。
基金资助:
国家自然科学基金(61702150)

Digital Image Source Forensics Based on Image Noise Residual

HUANG Mingying

School of Computer Science and Technology,Hangzhou Dianzi University,Hangzhou 310018,China

Received:2019-01-15 Online:2020-02-15 Published:2020-03-12
Supported by:
National Natural Foundation of China(61702150)

摘要/Abstract

摘要：

针对使用支持向量机的图像来源取证算法中存在的,例如所需训练集较大(数千幅)、特征维数较高等问题。文中提出了一种取证算法,该算法仅需要少量(约为10幅)训练图像,且只提取图像噪声的残差作为图像的唯一特征。该算法首先使用小波滤波器提取图像噪声,然后借助回归模型提取噪声的残差,最后为噪声的残差建立高斯分布模型,根据不同类型图像噪声的残差模型参数进行来源取证。实验结果表明,在误报率为1.2% 的条件下,该取证算法对自然图像的准确率为95.33%,对计算机生成图像的准确率为96.44%。

关键词: 自然图像, 计算机生成图像, 图像来源取证, 小波滤波器, 回归模型, 高斯模型

Abstract:

There are some problems with the image source forensics algorithm using the support vector machine, such as the large training set (about thousands) and the high dimension feature. To solve these problems, a forensic algorithm was proposed, which required only a small number (about 10) of training images, and only extracted the residual of the image noise as the unique feature of the image. The algorithm firstly used the wavelet filter to extract the image noise, and then extracted the residual of the image noise using the regression model. Finally, the Gaussian distribution model was established for the noise residual, and the source evidence was obtained according to different types of images with different model parameters. The experimental results showed that under the condition of FPR of 1.2%, the TPR of the proposed algorithm for natural images was 95.33%, and the TPR for computer-generated graphics was 96.44%.

Key words: natural images, computer-generated graphics, image source forensics, the wavelet filter, regression model, gaussian model

中图分类号:

TP319

黄明瑛. 基于图像噪声残差的数字图像来源取证[J]. 电子科技, 2020, 33(2): 37-42.

HUANG Mingying. Digital Image Source Forensics Based on Image Noise Residual[J]. Electronic Science and Technology, 2020, 33(2): 37-42.

图/表 6

图1

图2

图3

图4

表1

表2

参考文献 17

[1]	周琳娜, 王东明 , 数字图像取证技术[M]. 北京: 北京邮电大学出版社, 2008.
	Zhou Linna, Wang Dongming. Digital image forensics technology[M]. Beijing: Beijing University of Posts and Telecommunications Press, 2008.
[2]	Lyu S, Farid H . How realistic is photorealistc[J]. IEEE Transactions on Signal Processing, 2005,53(2):845-850.
[3]	Li J, Li X, Yang B , et al. Segmentation-based image copy-move forgery detection scheme[J]. IEEE Transactions on Information Forensics and Security. 2015,10(3):507-18.
[4]	Xia Z, Wang X, Zhang L , et al. A privacy-preserving and copy-deterrence content-based image retrieval scheme in cloud computing[J]. IEEE Transactions on Information Forensics and Security, 2016,11(11):2594-2608.
[5]	Zhou Z, Wang Y, Wu Q J , et al. Effective and efficient global context verification for image copy detection[J]. IEEE Transactions on Information Forensics and Security, 2017,12(1):48-63.
[6]	郑伟, 张大兴, 景阳帅 . 一种基于WLD特征的盲检测数字水印算法[J]. 电子科技, 2016,29(7):15-18.
	Zheng Wei, Zhang Daxing, Jing Yangshuai . A blind detection digital watermarking algorithm based on WLD features[J]. Electronic Science and Technology, 2016,29(7):15-18.
[7]	Gallagher A, Chen T . Image authentication by detecting traces of demosaicing[C]. Anchorage:IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, 2008.
[8]	Qiao T, Retraint F, Cogranne R . Image authentication by statistical analysis[C]. Marrakech:Proceedings of the 21 ^st European Signal Processing Conference , 2013.
[9]	Chang C C, Lin C J . LIBSVM:A library for support vector machines[J]. ACM Transactions on Intelligent Systems and Technology, 2011,2(3):1-27.
[10]	Wang J, Li T, Shi Y Q , et al. Forensics feature analysis in quaternion wavelet domain for distinguishing photographic images and computer graphics[J]. Multimedia Tools and Applications, 2016,76(22):1-17.
[11]	Peng F, Zhou D L, Long M , et al. Discrimination of natural images and computer generated graphics based on multi-fractal and regression analysis[J]. AEU-International Journal of Electronics and Communications, 2017,71(3):72-81.
[12]	Fridrich J, Lukas J, Goljan M . Digital camera identification from sensor noise[J]. IEEE Transactions on Information Security and Forensics, 2006,1(2):205-214.
[13]	Long M, Peng F, Zhu Y . Identifying natural images and computer generated graphics based on binary similarity measures of PRNU[J]. Multimedia Tools and Applications, 2017(7):1-18.
[14]	De Rezende E R, Ruppert G C, Carvalho T . Detecting computer generated images with deep convolutional neural networks[C]. Niteroi:The 30 ^th Conference on Graphics, Patterns and Images , 2017.
[15]	De Rezende E R, Ruppert G C, Theophilo A , et al. Exposing computer generated images by using deep convolutional neural networks[J]. Signal Processing:Image Communication, 2018(1):1352-1366.
[16]	Mader B, Banks M S, Farid H . Identifying computer-generated portraits: the importance of training and incentives[J]. Perception, 2017,46(9):1062-1076.
[17]	Gloet Bohme R . The dresden image database'for bench marking digital image forensics[C]. Sierre:Proceedings of the ACM Symposium on Applied Computing, 2010.

指标	本文算法	文献[2]	文献[11]	文献[8]	文献[7]
PNI	95.33%	98.78%	99.11%	56.90%	80.33%
PCG	96.44%	96.78%	97.89%	57.35%	73.44%

图像数量	0.1	0.2	0.3	0.4
n=10	97.00%	99.00%	99.56%	99.67%
n=20	97.11%	99.00%	99.56%	99.67%
n=30	97.22%	99.00%	99.56%	99.67%
n=40	97.22%	99.00%	99.56%	99.67%

基于图像噪声残差的数字图像来源取证

Digital Image Source Forensics Based on Image Noise Residual

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 17

相关文章 4

Metrics

本文评价

推荐阅读 10

[1]	张海兵. 基于粒子滤波的动态自回归预测模型方法[J]. , 2016, 29(6): 15-.
[2]	贾欣欣. 格兰杰因果关系在神经科学领域的发展及缺陷[J]. , 2015, 28(8): 178-.
[3]	杨姗姗,王松会,宋东东. 基于回归分析的研究及R语言实现[J]. , 2015, 28(10): 186-.
[4]	崔学超. 基于减背景和混合差分的运动目标检测算法[J]. , 2010, 23(10): 85-88.