基于变邻域局部搜索的篡改图像检测算法

The tampered image detection algorithm based on local search invariable neighborhood

针对复制篡改图像的多个篡改区域检测率较低的问题,提出了一种基于变邻域局部搜索的篡改图像检测算法。提取图像分块后的SIFT特征,通过特征匹配算法判断图像是否经过篡改,合并满足相似性条件的由分块的超像素所得到的疑似区域以及变邻域局部搜索算法所得到的邻域块,采用形态学闭运算填补孔洞,完成篡改区域的检测。改进算法的检测准确率为91.1%,相较于其他两种算法,检测准确率分别提高2.6%和0.6%。仿真结果表明,改进算法对于复制篡改图像的单个篡改区域和多个篡改区域检测效果较好。

Aiming at the problem of low detection rate of multiple tampered areas of copied tampered images,a tampered image detection algorithm based on variable neighborhood local search is proposed.Extract the SIFT features after the image is divided,and use the feature matching algorithm to determine whether the image has been tampered. The suspected regions obtained by the divided super-pixels and the neighborhood blocks obtained by the variable neighborhood local search algorithm that meet the similarity condition are merged. The morphological closed operation is used to fill the holes to complete the detection of the tampered area. The detection accuracy of the improved algorithm is 91.1%. Compared with the other two algorithms,the detection accuracy is increased by 2.6% and 0.6% respectively. The simulation results show that the improved algorithm has a better detection effect on the single tampered area and multiple tampered areas of the copied tampered image.