基于密度聚类的多重复制-粘贴篡改检测与定位算法

Multiple Copy-move Forgery Detection and Localization based on Density Clustering

为了克服当前篡改检测算法不能识别多重复制-粘贴篡改的问题,本文基于逆序广义2近邻和密度聚类方法,提出了一种能有效检测并定位多重复制-粘贴篡改图像的算法。实验表明,该算法时间复杂度较低,准确率较高,且能精确地识别多重复制-粘贴篡改并定位篡改区域。

The current tamper detection algorithm can not recognize multiple copy-move forgery.To solve this problem,this paper proposes an algorithm which can effectively detect and locate multiple copy-move forgery based on the reversed-generalized 2 nearest neighbor and density clustering method.Experiments show that the algorithm has low time complexity and high accuracy,and can accurately identify multiple copy-move forgery and locate tampered region.