基于SIFT的快速复制粘贴篡改检测算法  

Fast Copy Paste Tamper Detection Algorithm Based on SIFT

作　　者：刘欢欢 金冶纯 张军凯 LIU Huan-huan;JIN Ye-chun;ZHANG Jun-kai(Information Engineering Department,Hebei Institute of Mechanical and Electrical Technology,Xingtai,Hebei 054035,China;Department of Physics and Information Engineering,Cangzhou Normal University,Cangzhou,Hebei 061001,China;Department of Electrical Engineering,Hebei Institute of Mechanical and Electrical Technology,Xingtai,Hebei 054035,China)

机构地区：[1]河北机电职业技术学院信息工程系,河北邢台054035 [2]沧州师范学院物理与信息工程系,河北沧州061001 [3]河北机电职业技术学院电气工程系,河北邢台054035

出　　处：《沧州师范学院学报》2022年第1期49-53,102,共6页Journal of Cangzhou Normal University

摘　　要：针对传统图像复制粘贴篡改算法计算量大、耗时长的缺陷,提出一种基于SIFT图像特征点的快速篡改检测算法.首先利用SIFT算法针对图像的特征点进行检测,对于图像特征点的主方向以及描述子,采用Haar小波变换来确定;针对特征点维度较高需进行降维处理,采用PCA算法,最后使用优化的最近邻BBF搜索算法进行特征点的快速匹配,最终成功地检测出图像的复制粘贴篡改.Traditional image copy paste tamper algorithm takes large amount of calculation and long time.To solve the problem,a fast tamper detection algorithm based on SIFT is presented in this paper.Firstly,feature points are tested using SIFT algorithm.For each feature points,the main direction and feature point descriptor was ascertained by Haar wavelet transform.Then the high latitude of feature point was reduced by using the PCA algorithm.Finally,the fast matching of feature points is performed by using the nearest neighbor search algorithm BBF to detect copy paste tamper of the image.

关 键 词：数字图像 复制-粘贴篡改检测 SIFT算法 PCA BBF搜索算法 

分 类 号：TP37[自动化与计算机技术—计算机系统结构] TN919.81[自动化与计算机技术—计算机科学与技术]