顽健合成图像篡改检测及定位

doi:10.11959/j.issn.1000-0801.2017214

电信科学 ›› 2017, Vol. 33 ›› Issue (8): 120-129.doi: 10.11959/j.issn.1000-0801.2017214

顽健合成图像篡改检测及定位

王潇,张荣(),郭立君

宁波大学信息科学与工程学院，浙江宁波 315211

修回日期:2017-06-30 出版日期:2017-08-01 发布日期:2017-08-25
作者简介:王潇（1993-），女，宁波大学信息科学与工程学院硕士生，主要研究方向为数字图像取证与信息安全。|张荣（1974-），女，博士，宁波大学副教授，主要研究方向为数字取证与信息安全。|郭立君（1970-），男，博士，宁波大学教授，主要研究方向为计算机视觉与模式识别、移动互联网及其应用。
基金资助:
浙江省自然科学基金资助项目(LY17F030002);浙江省“信息与通信工程”重中之重学科开放基金资助项目(xkxl1521);浙江省“信息与通信工程”重中之重学科开放基金资助项目(xkxl1516)

Robust tampering detection and localization of composite image

Xiao WANG,Rong ZHANG(),Lijun GUO

College of Information Science and Engineering,Ningbo University,Ningbo 315211,China

Revised:2017-06-30 Online:2017-08-01 Published:2017-08-25
Supported by:
Zhejiang Provincial Natural Science Foundation of China(LY17F030002);Zhejiang Province “Information and Communication Engineering” in the Top Priority of the Subject of Open Fund(xkxl1521);Zhejiang Province “Information and Communication Engineering” in the Top Priority of the Subject of Open Fund(xkxl1516)

摘要/Abstract

摘要：

针对自然图像与高度仿真的计算机生成图像的合成图像篡改检测问题，提出在YCbCr颜色空间基于差分直方图和中心对称局部二进制模式提取图像块颜色和纹理特征的方法，通过训练后验概率支持向量机模型对待测图像块进行识别。在不重叠分块情况下先大致判断篡改区域，然后在该区域内逐像素分块判别，最终实现篡改区域精确定位。实验结果表明，对128 dpi × 128 dpi图像块的识别率达到94.75%，高于现有方法；对合成图像篡改区域能够实现精确定位，且对旋转、缩放操作表现出较好的顽健性。

关键词: 合成图像, 特征提取, 后验概率支持向量机, 篡改检测

Abstract:

Aiming at the problem of tamper detection of composite image of natural images and highly simulated computer-generated images,a method of extracting image block color and texture feature based on differential histogram and local binary texture descriptor in YCbCr color space was proposed.By training posterior probability support vector machine,the image block to be measured was identified.In the case of non-overlapping block,the approximate tampering area was general judged,then the block was discriminated by pixel in the region,ultimately the accurate location of tampering area was achieved.The experimental results show that the recognition rate of 128 dpi×128 dpi image blocks is 94.75%,which is higher than other methods.The tapering region of the synthesized image can be precisely positioned,and the rotation and scaling operation show good coercivity.

Key words: composite image, feature extraction, posterior probability support vector machine, tampering detection

中图分类号:

TP391

王潇,张荣,郭立君. 顽健合成图像篡改检测及定位[J]. 电信科学, 2017, 33(8): 120-129.

Xiao WANG,Rong ZHANG,Lijun GUO. Robust tampering detection and localization of composite image[J]. Telecommunications Science, 2017, 33(8): 120-129.

图/表 16

图1

图2

图3

图4

图5

图6

表1

图7

图8

图9

表2

图10

表3

表4

图11

参考文献 18

[1]	GAO S , ZHANG C , WU C L ,et al. A hybrid feature based method for distinguishing computer graphics and photo-graphic image[C]// International Workshop on Digital Watermarking,September 17-19,2013,Beijing,China. Berlin:Springer Berlin Heidelberg, 2013: 303-313.
[2]	HOLMES O , BANKS M S , FARID H . Assessing and improving the identification of computer-generated portraits[J].ACM Transactions on Applied Perception (TAP),2016,13(2):7. ACM Transactions on Applied Perception (TAP), 2016,13(2):7.
[3]	AYYALASOMAYAJULA R S , PANKAJAKSHAN V . Differentiating photographic and PRCG images using tampering localization features[C]// International Conference on Computer Vision and Image Processing,January 12-13,2016,Singapore.[S.l.:s.n]. 2017: 429-438.
[4]	王亚东, 张荣, 蒋桧慧 ,等. 引入垂直全局表观约束的中层特征行人再识别[J]. 电信科学, 2017,33(4): 114-125.
	WANG Y D , ZHANG R , JIANG H H ,et al. Person re-identification using mid-level features with vertical global appearance constraint[J]. Telecommunications Science, 2017,33(4): 114-125.
[5]	LYU S , FARID H . How realistic is photorealistic?[J]. IEEE Transactions on Signal Processing, 2005,53(2): 845-850.
[6]	PENG F , ZHOU D , 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): 72-81.
[7]	WU R , LI X , YANG B . Identifying computer generated graphics via histogram features[C]// 2011 18th IEEE International Conference on Image Processing (ICIP),September 11-14,2011,Brussels,Belgium. New Jersey:IEEE Press, 2011: 1933-1936.
[8]	LI Z , YE J , SHI Y Q . Distinguishing computer graphics from photographic images using local binary patterns[M]. Berlin: Springer Berlin HeidelbergPress, 2012: 228-241.
[9]	OJALA T , PIETIKAINEN M , MAENPAA T . Multi-resolution gray-scale and rotation invariant texture classification with local binary patterns[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002,24(7): 971-987.
[10]	申铉京, 李梦臻, 吕颖达 ,等. 基于 LBC 的计算机生成图像盲鉴别算法[J]. 计算机科学, 2015,42(6): 135-138.
	SHEN X J , LI M Z , LV Y D ,et al. Blind identification algorithm of photorealistic computer graphics based on local binary count[J]. Computer Science, 2015,42(6): 135-138.
[11]	CONOTTER V , CORDIN L . Detecting photographic and computer generated composites[C]// IS＆T/SPIE Electronic Imaging.International Society for Optics and Photonics,August 12-14,2011,Chennai,India.[S.l.:s.n. ], 2011: 3238-3241.
[12]	KE Y , MIN W , DU X ,et al. Detecting the composite of photographic image and computer generated image combing with color,texture and shape feature[J]. Journal of Theoretical ＆ Applied Information Technology, 2013,49(2).
[13]	CHANG C C , LIN C J . LIBSVM:a library for support vector machines[J]. ACM Transactions on Intelligent Systems and Technology(TIST), 2011,2(3)：27.
[14]	裴安山, 王让定, 严迪群 . 基于设备本底噪声频谱特征的手机来源识别[J]. 电信科学, 2017,33(1): 85-94.
	PEI A S , WANG R D , YAN D Q . Cell-phone origin identification based on spectral features of device self-noise[J]. Telecommunications Science, 2017,33(1): 85-94.
[15]	PLATT J . Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods[J]. Advances in Large Margin Classifiers, 2000,10(4): 61-74.
[16]	CHEN W . Detection of digital image and video forgeries[D]. Newark:New Jersey Institute of Technology, 2008.
[17]	M?ENP?? A , PIETIK? M . Texture analysis with local binary patterns[J]. Handbook of Pattern Recognition and Computer Vision, 2005(3): 197-216.
[18]	NG T T , CHANG S F , HSU J ,et al. Columbia photographic images and photorealistic computer graphics dataset[R]. 2005.

颜色空间	颜色通道	TPR	TNR	ACC
RGB	R	95%	91%	93%
	G	95.7%	91.2%	93.45%
	B	96.6%	90.6%	93.6%
HSV	H	96.3%	92.1%	94.2%
	S	95.2%	91.1%	93.15%
	V	97%	91%	94%
YCbCr	Y	96.5%	92.4%	94.45%
	Cb	97.2%	91.4%	94.3%
	Cr	96.3%	92.1%	94.2%

算法	块尺寸	特征维数	TPR	TNR	平均正确率
Wavelet^[5]	256 dpi × 256 dpi	216	84.2%	79.6%	81.9%
	128 dpi × 128 dpi		80.2%	78.8%	79.5%
Wavelet+SPN^[11]	256 dpi × 256 dpi	228	83.6%	88.6%	86.1%
	128 dpi × 128 dpi		80.8%	89.7%	85.25%
DH^[7]	256 dpi × 256 dpi	112	95.8%	92.9%	94.35%
	128 dpi × 128 dpi		95.6%	89.4%	92.5%
LBP^[8]	256 dpi × 256 dpi	236	93.6%	86.5%	90.05%
	128 dpi × 128 dpi		92.9%	84.1%	88.5%
DH+CSLBP	256 dpi × 256 dpi	144	98.9%	94.7%	96.8%
	128 dpi × 128 dpi		97.7%	91.8%	94.75%

检测率	0.6	0.8	1.0	1.2
TPR	78.5%	93.9%	94.5%	93.2%
FPR	22.3%	1.8%	3.3%	3.5%

检测率	45°	90°	135°	180°
TPR	78.6%	94.7%	94.2%	92.5%
FPR	12.6%	5.6%	8.9%	8.0%

顽健合成图像篡改检测及定位

Robust tampering detection and localization of composite image

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 16

参考文献 18

相关文章 9

Metrics

推荐阅读 0

[1]	尹良, 林睿, 王晓雷, 姚宇亮, 周林, 何元. 基于频谱形状的低复杂度雷达信号分类[J]. 电信科学, 2022, 38(1): 25-35.
[2]	叶学义, 钱丁炜, 应娜, 王涛. 基于吸引模式的局部二阶梯度轮廓人脸识别算法[J]. 电信科学, 2021, 37(7): 96-106.
[3]	蔡鑫. 基于Bert模型的互联网不良信息检测[J]. 电信科学, 2020, 36(11): 121-126.
[4]	朱琳,赵娟,王伊婷,冯俊兰,邓超. 移动通信网络投诉热点问题智能预警方法[J]. 电信科学, 2019, 35(5): 17-24.
[5]	薛栋,周亚训,金炜. 版权保护和内容认证的全盲双功能数字水印算法[J]. 电信科学, 2017, 33(2): 79-89.
[6]	陆寄远,黄承慧,侯昉,李斌. 基于Hadoop云计算平台的图像分类与标注[J]. 电信科学, 2014, 30(2): 65-69.
[7]	魏莎莎,陆慧娟,金伟,李超. 基于云平台的互信息最大化特征提取方法研究[J]. 电信科学, 2013, 29(10): 38-42.
[8]	闫云斌,全厚德,崔佩璋. 跳频通信中阻塞干扰样式的自动识别[J]. 电信科学, 2013, 29(1): 103-108.
[9]	闫云斌,全厚德,崔佩璋. 跳频通信中阻塞干扰样式的自动识别[J]. 电信科学, 2012, 28(12): 46-51.