【摘要】 提出了一种新的数字图像被动认证算法,用于检测同幅图像的区域复制窜改问题。算法首先利用离散小波变换提取图像的低频分量,再对低频分量进行分块并提取每一块的离散正交Tchebichef矩特征;然后将特征矢量进行字典排序,比较相邻两组特征矢量的相似性;最后利用阈值判别实现窜改伪造区域的检测和定位。实验结果证明,算法能较好地检测及定位出图像中复制与窜改区域,且具有运算量小、检测效率高、鲁棒性好等特点。更多还原

【Abstract】 This paper presented a new passive image authenticate algorithm which could detect the type of region duplication forgery in a single digital image.After extacting the image low-frequency by DWT,this algorithm applied the discrete orthogonal Tchebichef moments invariants to the fixed sized overlapping blocks of a low-frequency image in the wavelet sub-band.Then,it sorted the eigenvectors lexicographically,matched the similarity of neighboring eigenvectors,and identified the forgery part by the threshold finally.The experimental result verifies that the algorithm can detect the tampered region with the feature of small computational load,effective detection and good robustness.更多还原