基于SURF的图像多区域复制粘贴篡改检测  

Image multiple copy-move forgery detection based on SURF algorithm

作　　者：陈辉映 张大兴[1] 杨珊珊 郭家伟 CHEN Hui-ying, ZHANG Da xing, YANG Shan-shan, GUO Jia-wei(institute of image Graphics, Hangzhou Dianzi University, Hangzhou 310018, Chin)

机构地区：[1]杭州电子科技大学图形图像研究所,浙江杭州310018

出　　处：《计算机工程与设计》2018年第8期2593-2597,共5页Computer Engineering and Design

摘　　要：针对现有复制粘贴篡改检测算法对多区域复制粘贴篡改图像检测效果差、检测速度较慢的问题,提出一种基于SURF的图像多区域复制粘贴篡改检测方法。将待检测图像从RGB转换到Lab颜色空间并进行SURF特征提取,利用kg2NN方法对得到的特征点进行匹配,利用聚类消除误匹配定位篡改区域。实验结果表明,该算法能够准确定位图像的多处复制粘贴篡改区域,在抗几何变换攻击和JPEG压缩等方面具备一定的鲁棒性,检测速度得到了较大提高。Aiming at the problem that the performance of existing detection algorithm for multiple copy-move tampering image detection is poor and the detection speed of it is slow,a method of image copy and paste tamper detection based on SURF algorithm was proposed.The image to be detected was transformed from the RGB color space to the Lab color space and the SURF feature was extracted.The obtained feature points were matched using k-g2 NN method.The clustering was used to eliminate the misalignment to locate the tampering area.Experimental results show that the proposed algorithm can accurately detect the multiple copy-move tampering,and it has certain robustness to anti-geometric transformation attack,and the detection speed is greatly improved.

关 键 词：特征点 加速鲁棒特征 鲁棒性 篡改检测 复制粘贴 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]