计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (18): 156-161.

• 图形图像处理 • 上一篇    下一篇

数字图像模式噪声篡改反取证

杨  弘,周治平   

  1. 江南大学 物联网工程学院,江苏 无锡 214122
  • 出版日期:2014-09-15 发布日期:2014-09-12

Anti-forensic research of image sensor-pattern noise forger

YANG Hong, ZHOU Zhiping   

  1. College of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2014-09-15 Published:2014-09-12

摘要: 针对基于模式噪声的图像来源性取证技术,提出一种全新的误导攻击方式,分析了特定相机的相机成像模型,推导出了模式噪声替换公式,该公式抑制了原有相机的模式噪声痕迹,并附加上了伪造相机的模式噪声,通过线性回归或基于图像内容特征的方式求出合理的嵌入参数。实验表明,该算法不但可以有效地混淆来源性取证,更可以让伪造者控制取证技术的结论,从而达到误导的效果。

关键词: 数字图像, 来源性取证, 反取证, 模式噪声, 非均一响应, 回归分析

Abstract: Aiming at the image source identification forensic technology based on pattern noise, this paper proposes a new mislead attack method, it analyses the sensor out model of the specific camera, deduces the pattern noise substitution formula. This formula can restrain the origin pattern noise and add the fake camera’s noise. By the regression analysis or figure, it can get a proper embedding parameter. Experiments show that the algorithm can not only confuse forensic technology, but also lead the tester to the wrong way.

Key words: digital image, source forensic, anti-forensic, pattern-noise, photo-response non uniformity, regression analysis