基于改进SIFT算法的图像复制粘贴篡改检测

Image Copy-Paste Tampering Detection Based on Improved SIFT Algorithm

复制粘贴是一种常见的图像篡改方式,也是最隐蔽的图像篡改手段之一。SIFT是一种常用的匹配算法,同时也是一种较为有效的复制粘贴图像篡改检测方法。目前基于SIFT的图像篡改检测方法中,存在着匹配精度差及时间复杂度高等问题。为了克服这些问题,对SIFT算法进行了改进:针对阈值增大造成精确性差的问题,采用拟合优化的方法确定阈值,对SIFT算法中提取特征点的方法进行了改进;针对SIFT算法特征匹配阶段时间复杂度高的问题,采用基于K-D树的BBF搜索算法进行最近邻查询以实现特征点的快速匹配,对SIFT算法中的特征匹配进行了改进。实验结果验证了该算法的有效性。

Copy-paste is a technique widespreadly used in image tampering,and it is one of the most covert tampering means.SIFT is a kind of common matching algorithm,it is also a more effective test method for the copy-paste tampering images.But it has the problems of poor matching accuracy,high time-complexity and so on.In order to overcome the problems,some improvements were made in this paper.The threshold value is determined by the method of fitting optimization to solve the problem of accuracy when the threshold is increased,and methods of extracting the feature points by SIFT algorithm are improved.The BBF search algorithm based on K-D tree is adopted to achieve the fast matching of the nearest neighbor query,and the feature matching is improved in SIFT algorithm,so the problem of high time complexity goes easy.The experimental results show that the proposed algorithms are effective.