图像源辨识反攻击算法的简化与分析  

Simplification and Analysis of Counter-Attack Algorithm for Image Source Identification

作　　者：胡永健[1] 关华[1] 赵静[1] 刘琲贝[1] 

机构地区：[1]华南理工大学电子与信息学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2014年第7期1-8,共8页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家"973"计划项目(2011CB707003);国家自然科学基金专项基金项目(61327005);华南理工大学中央高校基本科研业务费专项资金资助项目(2014ZZ0036)

摘　　要：基于相机指纹/成像传感器模式噪声的图像源辨识算法容易受到指纹复制攻击.文中首先对现有的三角测试反攻击算法进行了简化,即通过将图像各块的衰减因子近似认为是常数,减少了计算时间,然后从传感器模式噪声添加方式、图像内容、模式噪声提取方式和盗用图片数目四个方面对反攻击算法性能进行了详细分析.最后,根据实验结果指出,以上四个方面都会影响反攻击算法的成功率.文中还讨论了这些因素所影响的程度和范围.Considering the image source identification algorithm based on the camera fingerprint and the imaging sen -sor pattern noise to be easily attacked by fingerprint-copy operation , this paper simplifies the triangle test algorithm which is a counter-attack algorithm against the fingerprint-copy operation , that is, the attenuation factor for each block is set as a constant , thus reducing calculation time .Then, the performance of the counter-attack algorithm is discussed in detail from four aspects , namely, the addition manner of sensor pattern noise , the image content , the way of extracting sensor pattern noise and the number of illegally -used images.Finally, according to the experimental results, this paper points out that all the four aspects can affect the success rate of the counter -attack algorithm .In addition, the degree and scope of effect caused by the above-mentioned factors are also explored .

关 键 词：相机 数字图像取证 图像源辨识 传感器模式噪声 反取证 反-反取证 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]