基于SIFT的数字图像盲检测算法研究  

Based on the SIFT digital image blind detection algorithms

作　　者：徐亮[1] 魏锐[1] 

机构地区：[1]黄淮学院信息工程学院,河南驻马店463000

出　　处：《电子设计工程》2013年第12期117-120,共4页Electronic Design Engineering

基　　金：河南省重点科技攻关项目(112102210335)

摘　　要：针对图像区域的复制-变换-移动-粘贴篡改,建立了相应的篡改模型,提出了基于SIFT标记图特征向量的盲鉴别算法,能够有效地检测并定位经过旋转、缩放变换的篡改区域,而且能够抵抗高斯模糊、高斯白噪声和JPEG重压缩等后处理。与基于传统SIFT特征向量的检测算法相比,该算法对篡改区域的缩放变换具有更高的鲁棒性。In image area replication-transformation-mobile-paste manipulation, establishes the tamper with the model, is proposed based on SIFT mark image feature vector of the blind identification algorithm, can effectively detect and positioning after rotation, scaling transformation of tamper with the area, but also can resist gaussian blur, gaussian white noise and JPEG compression heavy post-processing. And based on the traditional SIFT feature vector detection algorithm, this algorithm to tamper with the regional scale transformation has higher robustness.

关 键 词：图像篡改 特征向量 数字图像盲鉴别 SIFT算法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]