基于伪Zernike矩的数字图像篡改检测方法  

A Scheme of Digital Image Copy-move Forgery Detection Based on Pseudo-Zernike Moments

作　　者：唐腾健[1] 朱玥[2] TANG Tengjian;ZHU Yue(school of Artificial Intelligence and Big Data,Chongqing Industry Polytechnic College,Chongqing 401120,China;School of Design,Chongqing Industry Polytechnic College,Chongqing 401120,China)

机构地区：[1]重庆工业职业技术学院人工智能与大数据学院,重庆401120 [2]重庆工业职业技术学院设计学院,重庆401120

出　　处：《长江信息通信》2024年第10期30-33,共4页Changjiang Information & Communications

摘　　要：图像编辑技术的快速发展使得人们难以发现专业技术处理过的篡改图像的伪造痕迹,对新闻传播、司法取证、信息安全等领域造成了严重的威胁。图像的区域篡改检测方案主要用于检测图像中存在疑似相似的图像块。目前,传统的检测策略对发生几何攻击的篡改区域鲁棒性不好。针对几何攻击情况,文章介绍了一种针对复制粘贴伪造图像的新检测方法,利用伪Zernike矩特征进行特征提取。通过在相邻区域进行特征匹配,可以检测出图像中可能存在的伪造区域。实验结果表明,该方法不仅对常规攻击(如加噪声、JPEG压缩)具有较好的检测效果,而且在处理旋转等几何攻击时也表现出色。这些结果证明了该方法的创新性和广泛适用性。The rapid development of image cditing technology has made it difficult to detct traces of forgery in professionally processed tampered images,posing a serious threat to the ficlds of news dissemination,judicial forensics,and information sccurity.Region tampering de-tection schemes for images are mainly used to detect the presence of suspected similar image blocks in an image.Currently,traditional detection strategies are not robust to tampered regions where geometric attacks occur.For the gcomctric attack scenario,this paper introduces a new de-tection method for copy-paste forged images,which utilizes pseudo-Zernike moment features for feature extraction.By matching the features in neighboring regions,the possible forged re-gions in the image can be detected.Experimcntal results show that the method not only has good detection effect on conventional attacks(e.g.,add noise,JPEG compression),but also performs well in dealing with geometric attacks such as rotation.These results demonstrate the innovation and wide applicability of the method.

关 键 词：几何攻击 伪ZERNIKE矩 复制粘贴 信号处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]