基于传感器模式噪声特性的图像篡改检测方法

Image forgery detection method based on sensor pattern noise characteristics

针对传统基于传感器模式噪声特性的图像篡改检测算法由于需要知道参考图像数据库因而应用局限性大的问题,提出了一种基于噪声子空间投影的图像篡改检测框架,分别采用主成分分析(PCA)、二维主成分分析(2DPCA)和核主成分分析(KPCA)实现了基于图像噪声特性的篡改检测,并通过实验验证了此方法的有效性。

Aiming at problem that traditional characteristics image forgery detecting algorithm based on sensor pattern noise require reference image database,which restricts their practicable application,an image forgery detecting framework based on noise sub-space projection is proposed,principal component analysis（ PCA）,twodimensional PCA（ 2DPCA） or kernel PCA（ KPCA） are used,to realize forgery detection based on image noise characteristics effectiveness of the methods is illustrated by experiment.