论文部分内容阅读
针对现有数字图像复制-粘贴篡改检测中尺度不变特征变换(Scale Invariant Feature Transform,简称SIFT)算法计算复杂度高的问题,文章提出一种将SIFT特征点和中心对称局部二进制模式(Center Symmet-ric-Local Binary Pattern,简称CS-LBP)描述子相结合的篡改检测方法.首先提取SIFT关键点,再对每个关键点生成CS-LBP特征描述子,并利用K-D(k-dimensional)树和BBF(Best-Bin-First)搜索算法寻找符合特征匹配关系的匹配点对,判断是否存在图像区域的篡改.实验表明,与同类算法相比,所提出算法在不损失检测精度的同时有效地减少了运算量.“,”The existing scale invariant feature transform(SIFT)-based detection algorithm for copy-paste forgery has high computational complexity.To solve this problem,a new algorithm using the SIFT key-points and center symmetric-local binary pattern(CS-LBP) descriptor is proposed.Firstly,the SIFT key-points are determined by using the SIFT algorithm.Secondly,the CS-LBP descriptor for each key-point is generated.Finally,the k-dimensional(K-D) tree and the best-bin-first(BBF) search algorithm are used to locate the matched pairs of key-points and to determine whether there is any copy-paste forgery in the image.Experimental results show that compared with other algorithms,the proposed algorithm remarkably reduces the computational load under the same detection accuracy.