视频对象移除篡改的时空域定位被动取证

doi:10.11959/j.issn.1000-436x.2020151

通信学报 ›› 2020, Vol. 41 ›› Issue (7): 110-120.doi: 10.11959/j.issn.1000-436x.2020151

视频对象移除篡改的时空域定位被动取证

陈临强¹,杨全鑫²,袁理锋¹,姚晔¹(),张祯¹,吴国华¹

¹ 杭州电子科技大学网络空间安全学院，浙江杭州 310018
² 杭州电子科技大学计算机学院，浙江杭州 310018

修回日期:2020-04-29 出版日期:2020-07-25 发布日期:2020-08-01
作者简介:陈临强（1963- ），男，浙江临海人，杭州电子科技大学教授，主要研究方向为计算机图形学、视频实时处理、图形图像处理、定密理论与实务|杨全鑫（1994- ），男，河南林州人，杭州电子科技大学硕士生，主要研究方向为多媒体内容安全、图形图像处理|袁理锋（1983- ），男，浙江诸暨人，博士，杭州电子科技大学讲师，主要研究方向为图像内容安全、视觉秘密分享|姚晔（1978- ），男，湖北随州人，博士，杭州电子科技大学副教授，主要研究方向为多媒体内容安全、视频图像智能分析|张祯（1978- ），男，山东大同人，博士，杭州电子科技大学副教授，主要研究方向为计算机应用、保密信息化、图形图像处理|吴国华（1970- ），男，山东济南人，博士，杭州电子科技大学教授、博士生导师，主要研究方向为保密信息化、定密理论与实务
基金资助:
教育部人文社科基金资助项目(17YJC870021)

Passive forensic based on spatio-temporal localization of video object removal tampering

Linqiang CHEN¹,Quanxin YANG²,Lifeng YUAN¹,Ye YAO¹(),Zhen ZHANG¹,Guohua WU¹

¹ School of Cyberspace Security,Hangzhou Dianzi University,Hangzhou 310018,China
² School of Computer,Hangzhou Dianzi University,Hangzhou 310018,China

Revised:2020-04-29 Online:2020-07-25 Published:2020-08-01
Supported by:
Humanities and Social Sciences Foundation of Ministry of Education of China(17YJC870021)

摘要/Abstract

摘要：

针对视频被动取证领域中视频内容的真实性和完整性鉴定及篡改区域定位问题，提出了一种基于视频噪声流的深度学习检测算法。首先，构建了基于空间富模型（SRM）和三维卷积（C3D）神经网络的特征提取器、帧鉴别器和基于区域建议网络（RPN）思想的空域定位器；其次，将特征提取器分别与帧鉴别器和空域定位器相结合，搭建出2个神经网络；最后，利用增强处理后的数据训练出2种深度学习模型，分别用于对视频篡改区域时域和空域的定位。测试结果表明，时域定位的准确率提高到98.5%，空域定位与篡改区域标注平均交并比达49%，可以有效对该类篡改视频进行篡改区域时空域定位。

关键词: 视频对象移除篡改, 时空域定位, 视频被动取证, 三维卷积目标检测

Abstract:

To address the problem of identification of authenticity and integrity of video content and the location of video tampering area,a deep learning detection algorithm based on video noise flow was proposed.Firstly,based on SRM (spatial rich model) and C3D (3D convolution) neural network,a feature extractor,a frame discriminator and a RPN (region proposal network) based spatial locator were constructed.Secondly,the feature extractor was combined with the frame discriminator and the spatial locator respectively,and then two neural networks were built.Finally,two kinds of deep learning models were trained by the enhanced data,which were used to locate the tampered area in temporal domain and spatial domain respectively.The test results show that the accuracy of temporal-domain location is increased to 98.5%,and the average intersection over union of spatial localization and tamper area labeling is 49%,which can effectively locate the tamper area in temporal domain and spatial domain.

Key words: video object removal tampering, spatio-temporal localization, video passive forensic, object detection based on 3D convolution

中图分类号:

TP309

陈临强,杨全鑫,袁理锋,姚晔,张祯,吴国华. 视频对象移除篡改的时空域定位被动取证[J]. 通信学报, 2020, 41(7): 110-120.

Linqiang CHEN,Quanxin YANG,Lifeng YUAN,Ye YAO,Zhen ZHANG,Guohua WU. Passive forensic based on spatio-temporal localization of video object removal tampering[J]. Journal on Communications, 2020, 41(7): 110-120.

图/表 14

图1

图2

图3

图4

图5

图6

图7

图8

表1

表2

图9

表3

图10

表4

参考文献 19

[19]	SONG W , REN D , YU J ,et al. A new special video detection algorithm based on 3D convolutional co-occurrence gradient histogram and multi-instance learning[J]. Chinese Journal of Computers, 2019,42(1): 151-165.
[20]	HUANG S Y , RAMANAN D . Expecting the unexpected:training detectors for unusual pedestrians with adversarial imposters[C]// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2017: 4664-4673.
[21]	王青, 张荣 . 基于DCT系数双量化映射关系的图像盲取证算法[J]. 电子与信息学报, 2014,36(9): 2068-2074.
	WANG Q , ZHANG R . A blind Image forensic algorithm based on double quantization mapping relationship of DCT coefficients[J]. Journal of Electronics ＆ Information Technology, 2014,36(9): 2068-2074.
[22]	武伟, 詹玲超 . 利用颜色滤波阵列特性和模糊估计检测篡改[J]. 计算机工程与设计, 2007,28(21): 5179-5180.
	WU W , ZHAN L C . Detection of tampering using color filter array characteristics and fuzzy estimation[J]. Computer Engineering and Design, 2007,28(21): 5179-5180.
[23]	张静, 陈静, 苏育挺 . 基于滤波检测的视频区域篡改检测算法[J]. 电子测量技术, 2011,34(11): 66-69.
	ZHANG J , CHEN J , SU Y T . Detection of region-duplication forgery in the video streams[J]. Electronic Measurement Technology, 2011,34(11): 66-69.
[24]	杨弘, 周治平, 周翠娟 ,等. 基于模式噪声的手机图像篡改检测[J]. 计算机系统应用, 2013,22(9): 210-213.
	YANG H , ZHOU Z P , ZHOU C J ,et al. Mobile image tampering detection based on pattern noise[J]. Journal of Computer System Applications, 2013,22(9): 210-213.
[25]	王鑫, 鲁志波 . 基于JPEG块效应差异的图像篡改区域自动定位[J]. 计算机科学, 2010,37(2): 269-273.
	WANG X , LU Z B . Automatic localization of image tampering area based on JPEG block effect difference[J]. Computer Science, 2010,37(2): 269-273.
[1]	骆伟祺, 黄继武, 丘国平 . 鲁棒的区域复制图像篡改检测技术[J]. 计算机学报, 2007,30(11): 112-121.
	LUO W Q , HUANG J W , QIU G P . Robust tamper detection technology of regional reproduction image[J]. Chinese Journal of Computers, 2007,30(11): 112-121.
[26]	王传旭, 张祥光, 原春锋 ,等. 基于邻域相关性和帧间连续性的前景目标分割[J]. 数据采集与处理, 2007,22(3): 288-291.
	WANG C X , ZHANG X G , YUAN C F ,et al. Foreground target segmentation based on neighborhood correlation and frame continuity[J]. Data Collection and Processing, 2007,22(3): 288-291.
[2]	胡永健, AL-HAMIDI S, 王宇飞, ,等. 视频篡改检测数据库的构建及测试[J]. 华南理工大学学报(自然科学版), 2017,45(12): 57-64.
	HU Y J , AL-HAMIDI S , WANG Y F ,et al. Construction and testing of video tamper detection database[J]. Journal of South China University of Technology (Natural Science Edition), 2017,45(12): 57-64.
[27]	阳婷, 官洪运 . 基于帧间高频能量和相关性的烟雾检测算法研究[J]. 微型机与应用, 2015(17): 36-39.
	YANG T , GUAN H Y . Research on smoke detection algorithm based on high frequency energy and correlation between frames[J]. Microcomputer and Application, 2015(17): 36-39.
[3]	陈威兵, 杨高波, 陈日超 ,等. 数字视频真实性和来源的被动取证[J]. 通信学报, 2011,32(6): 177-183.
	CHEN W B , YANG G B , CHEN R C ,et al. Digital video passive forensics for its authenticity and source[J]. Journal on Communications, 2011,32(6): 177-183.
[4]	YAO Y , SHI Y Q , WENG S W ,et al. Deep learning for detection of object-based forgery in advanced video[J]. Symmetry, 2017,10(1):3.
[5]	姚晔, 胡伟通, 任一支 ,等. 数字视频区域篡改的检测与定位[J]. 中国图象图形学报, 2018(6): 779-791.
[28]	DU T , LUBOMIR B , ROB F ,et al. Learning Spatiotemporal Features with 3D Convolutional Networks[C]// Proceedings of 2015 IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2015: 4489-4497.
[29]	CHEN S D , TAN S Q , LI B ,et al. Automatic detection of object-based forgery in advanced video[J]. IEEE Transactions on Circuits ＆ Systems for Video Technology, 2016,26(11): 2138-2151.
[5]	YAO Y , HU W T , REN Y Z ,et al. Detection and location of digital video region tampering[J]. Journal of Image and Graphics, 2018(6): 779-791.
[6]	李岩, 刘念, 张斌 ,等. 图像镜像复制粘贴篡改检测中的 FI-SURF算法[J]. 通信学报, 2015,36(5): 54-65.
	LI Y , LIU N , ZHANG B ,et al. FI-SURF algorithm in image mirror copy and paste tamper detection[J]. Journal on Communications, 2015,36(5): 54-65.
[7]	黄维隽, 王士林 . 一种基于优化马尔可夫特征的图像篡改盲检测算法[J]. 信息安全与通信保密, 2014(3): 93-98,103.
	HUANG W J , WANG S L . An image tampering blind detection algorithm based on optimized Markov features[J]. Information Security and Communication Security, 2014(3): 93-98,103.
[8]	文伟, 肖志云, 彭思龙 . 边缘指导的BDCT压缩图像的后处理算法[J]. 计算机辅助设计与图形学学报, 2005(9): 2022-2028.
	WEN W , XIAO Z Y , PENG S L . Post processing algorithm of BDCT compressed image guided by edge[J]. Journal of Computer Aided Design and Graphics, 2005(9): 2022-2028.
[9]	耿庆田, 赵宏伟 . 基于分形维数和隐马尔可夫特征的车牌识别[J]. 光学精密工程, 2013(12): 216-222.
	GENG Q T , ZHAO H W . License plate recognition based on fractal dimension and hidden Markov features[J]. Optical Precision Engineering, 2013(12): 216-222.
[10]	ZHOU P , HAN X T , MORARIU V I ,et al. Learning rich features for image manipulation detection[C]// Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2018: 1053-1061.
[11]	REN S Q , HE K M , GIRSHICK R ,et al. Faster R-CNN:towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis ＆ Machine Intelligence, 2015,39(6): 1137-1149.
[12]	桑军, 郭沛, 项志立 ,等. Faster-RCNN的车型识别分析[J]. 重庆大学学报, 2017,40(7): 32-36.
	SANG J , GUO P , XIANG Z L ,et al. Vehicle detection based on fast er-RCNN[J]. Journal of Chongqing University, 2017,40(7): 32-36.
[13]	FAN Q F , BROWN L , SMITH J . A closer look at Faster R-CNN for vehicle detection[C]// Proceedings of 2016 IEEE Intelligent Vehicles Symposium (IV). Piscataway:IEEE Press, 2016: 124-129.
[14]	刘雨青, 黄添强 . 基于时空域能量可疑度的视频篡改检测与篡改区域定位[J]. 南京大学学报(自然科学), 2014(1): 61-71.
	LIU Y Q , HUANG T Q . Video tamper detection and tamper location based on energy suspicious degree in space-time domain[J]. Journal of Nanjing University (Natural Science), 2014(1): 61-71.
[15]	李倩, 王让定, 徐达文 . 基于视频修复的运动目标删除篡改行为的检测算法[J]. 光电子·激光, 2016,27(2): 182-190.
	LI Q , WANG R D , XU D W . Detection to video moving object deletion based on video inpainting[J]. Optoelectronics · Laser, 2016,27(2): 182-190.
[16]	何沛松 . 基于重编码痕迹的数字视频被动取证算法研究[D]. 上海:上海交通大学, 2018.
	HE P S . Research on digital video passive forensics algorithm based on recoded traces[D]. Shanghai:Shanghai Jiaotong University, 2018.
[17]	MOLCHANOV P , GUPTA S , KIM K ,et al. Hand gesture recognition with 3D convolutional neural networks[C]// Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway:IEEE Press, 2015: 1-7.
[18]	MOLCHANOV P , YANG X D , GUPTA S ,et al. Online detection and classification of dynamic hand gestures with recurrent 3D convolutional neural networks[C]// Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 4207-4215.
[30]	JIANG B R , LUO R X , MAO J Y ,et al. Acquisition of localization confidence for accurate object detection[C]// European Conference on Computer Vision. Berlin:Springer, 2018: 816-832.
[19]	宋伟, 任栋, 于京 ,等. 一种新的基于三维卷积共生梯度直方图和多示例学习的特殊视频检测算法[J]. 计算机学报, 2019,42(1): 151-165.

不同算法	测试帧数	Suc_rate	IOU_mean
VGG16 + RPN	4 557	68.61%	40.88%
SRM + VGG16 + RPN	4 557	89.99%	45.65%
C3D神经网络 + RPN	4 594	46.52%	29.08%
SRM + C3D神经网络 + RPN（本文算法）	4 594	94.47%	49.07%

输入	检测帧数	Suc_rate	IOU_mean
单帧	4 594	75.72%	40.40%
连续2帧	4 594	91.14%	46.81%
连续3帧	4 594	93.29%	47.94%
连续4帧	4 594	94.14%	49.28%
连续5帧	4 594	94.47%	49.07%

输出决策方案	检测帧数	Suc_rate	IOU_mean
方案a	4 594	94.47%	49.07%
方案b	4 594	96.93%	43.96%
方案c	4 594	98.11%	37.81%

测试视频	总帧数	实际篡改帧数	预测篡改帧数	预测正确的篡改帧数	FACC	Suc_rate	IOU_mean
视频1	285	123	108	101	89.82%	87.12%	48.64%
视频2	284	117	116	116	99.65%	100.0%	68.20%
视频3	292	139	134	134	98.29%	99.25%	50.77%
视频4	284	120	118	118	99.29%	94.06%	34.77%
视频5	291	103	77	76	90.38%	97.36%	46.27%
视频6	284	71	67	67	98.59%	97.01%	37.27%
视频7	291	98	98	97	99.31%	97.93%	71.45%
视频8	292	64	63	61	98.28%	93.44%	38.41%
视频9	291	180	182	179	98.62%	98.32%	57.26%
视频10	292	130	132	130	99.31%	95.38%	53.27%

视频对象移除篡改的时空域定位被动取证

Passive forensic based on spatio-temporal localization of video object removal tampering

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 19

相关文章 15

Metrics

推荐阅读 0

[1]	王文娟, 杜学绘, 单棣斌. 基于动态概率攻击图的云环境攻击场景构建方法[J]. 通信学报, 2021, 42(1): 1-17.
[2]	田苗苗, 陈静, 仲红. 格上基于身份的增量签名方案[J]. 通信学报, 2021, 42(1): 108-117.
[3]	江沛佩, 王骞, 陈艳姣, 李琦, 沈超. 区块链网络安全保障：攻击与防御[J]. 通信学报, 2021, 42(1): 151-162.
[4]	刘敖迪, 杜学绘, 王娜, 乔蕊. 基于深度学习的ABAC访问控制策略自动化生成技术[J]. 通信学报, 2020, 41(12): 8-20.
[5]	于斌, 黄海, 刘志伟, 赵石磊, 那宁. 面向多椭圆曲线的高速标量乘法器设计与实现[J]. 通信学报, 2020, 41(12): 100-109.
[6]	张兴兰,尹晟霖. 可变融合的随机注意力胶囊网络入侵检测模型[J]. 通信学报, 2020, 41(11): 160-168.
[7]	熊金波,毕仁万,陈前昕,刘西蒙. 边缘协作的轻量级安全区域建议网络[J]. 通信学报, 2020, 41(10): 188-201.
[8]	王佳琪,鲁宁,程庆丰,巫朝霞,史闻博. 面向隐私保护的多属性逆向频谱拍卖方案[J]. 通信学报, 2020, 41(9): 77-91.
[9]	赖英旭,薄尊旭,刘静. 基于改进PBFT算法防御区块链中sybil攻击的研究[J]. 通信学报, 2020, 41(9): 104-117.
[10]	田有亮,骆琴. 基于改进Merkle-Tree认证方法的可验证多关键词搜索方案[J]. 通信学报, 2020, 41(9): 118-129.
[11]	韩培义,刘川意,王佳慧,段少明,潘鹤中,方滨兴. 面向云存储的数据加密系统与技术研究[J]. 通信学报, 2020, 41(8): 55-65.
[12]	周由胜,谭畅,唐飞. 面向雾增强型工业物联网的多维安全查询方案[J]. 通信学报, 2020, 41(8): 175-186.
[13]	牛淑芬,刘文科,陈俐霞,王彩芬,杜小妮. 基于联盟链的可搜索加密电子病历数据共享方案[J]. 通信学报, 2020, 41(8): 204-214.
[14]	王娜,郑坤,付俊松,李剑. 基于分块的移动边缘计算密文检索方法[J]. 通信学报, 2020, 41(7): 95-102.
[15]	田有亮,吴雨龙,李秋贤. 基于信息论的入侵检测最佳响应方案[J]. 通信学报, 2020, 41(7): 121-130.