Abstract:
For the type of problem of region duplication forgery in the digital image, a passive algorithm of digital image, based on its own SIFT feature point, is put forward, which can resist geometric transformation. It extracts the image feature points with SIFT algorithm, and matches them. According to the feature that the feature point does not match one another in the same natural image, it can detect the tampered region which is translation, rotation, scaling and other geometric transformations. Experiment verifies the effectiveness of the algorithm of digital image region duplication forgery, which can resist geometric transformation with SIFT feature point.
Key words:
passive authentication,
region duplication forgery,
SIFT algorithm,
feature point,
anti-geometric transformation
摘要: 针对同幅图像的区域复制篡改问题,提出一种基于SIFT特征点的抗几何变换数字图像被动认证算法。在利用SIFT算法提取出图像中的SIFT特征点后,对特征点进行匹配。根据同一幅自然图像不会存在互相匹配特征点的这一特性,可以检测出篡改图像中平移、旋转、缩放等几何变换的区域。实验结果证明,该算法能够对抗区域复制篡改的几何变换。
关键词:
被动认证,
区域复制篡改,
SIFT算法,
特征点,
抗几何变换
CLC Number:
GU Zong-Yun, LV Huan-Li, LUO Bin, HAN Cheng-Mei. Image Passive Authentication Algorithm of Anti-geometric Transformation Based on Feature Point[J]. Computer Engineering, 2012, 38(2): 229-230.
谷宗运, 吕皖丽, 罗斌, 韩成美. 基于特征点的抗几何变换图像被动认证算法[J]. 计算机工程, 2012, 38(2): 229-230.