抗翻转、旋转和缩放攻击的图像区域复制篡改检测  

Detection of Image Region-Duplication Forgery Affected by Flipping, Rotation and Scaling

作　　者：邵虹[1] 朱虹[1] 崔文成[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870

出　　处：《计算机辅助设计与图形学学报》2015年第1期157-165,共9页Journal of Computer-Aided Design & Computer Graphics

基　　金：辽宁省自然科学基金(201202162);辽宁省高等学校优秀人才支持计划(LJQ2013013)

摘　　要：针对现有图像区域复制篡改检测算法对几何攻击较脆弱的特点,提出一种能够有效抵抗翻转、旋转和缩放的篡改检测方法.首先采用对数极坐标变换将圆窗图像投影到具有旋转、缩放不变性的坐标空间;然后自适应调整圆窗大小与频带宽度,消除多余图像信息与噪声的干扰,提高相位相关峰值以实现图像圆窗匹配;最后采用自适应调整得到的圆窗图像计算相位相关矩阵峰值位置以获取旋转角度与缩放因子,并将其应用于区域搜索,最终实现被复制区域与篡改区域的定位.实验结果表明,该方法对常用的几何变换操作准确率高、误检率低.Most existing region-duplication forgery detection algorithms are vulnerable to geometric distortions.Therefore, an effective tampering detection algorithm is proposed for resisting to geometric distortionswhich involves flipping, rotation, and scaling. Firstly, the circle window images are projected to the rotationand scaling invariance coordinate space by performing log-polar transform. Then circle window radius andfrequency band width can be adjusted automatically, which can eliminate redundant image information andnoise disturbance, and also increase the phase correlation peak in order to match the circle window images.Finally, the phase correlation can be calculated by adopting automatic adjustment circle window images toobtain the peak position, thus rotation angle and scaling factor can be got. These offsets are applied in regionsearching, and copied regions and tampered regions can be located. Experimental results show that the proposedmethod can detect region-duplication forgery image with common geometric distortions, which hashigh accuracy rate and low false detection rate.

关 键 词：篡改检测 区域复制 几何攻击 对数极坐标变换 相位相关 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]