【源码】基于SIFT和RANSAC算法的高分辨率图像伪造检测

克隆（复制-移动伪造）是对数字图像的恶意篡改攻击，其中图像的一部分被复制并粘贴到该图像中，以隐藏图像的重要细节，而不存在任何明显的操作痕迹。

Cloning (copy-move forgery) is a malicious tampering attack with digital images where a part of image is copied and pasted within the image to conceal the important details of image without any obvious traces of manipulation.

这种类型的篡改攻击给取证留下了图像真实性的大问题。

This type of tampering attacks leaves a big question of authenticity of images to the forensics.

在开发出强大的图像处理软件之后，过去的几年中已经提出了许多技术。

Many techniques are proposed in the past few years after powerful software’s are developed to manipulate the image.

该方案结合了基于块和基于特征点的提取技术，可以更准确地提取伪造区域。

The proposed scheme is involved with both the block based and feature point extraction based techniques to extract the forged regions more accurately.

该算法主要通过计算单位向量间的点积来匹配从每个块中提取的相同特征。

The proposed algorithm mainly involves in matching the tentacles of same features extracted from each block by computing the dot product between the unit vectors.

采用随机样本一致性（RANSAC）算法提取匹配区域。

Random Sample Consensus (RANSAC) algorithm is used to extract the matched regions.

实验结果表明，与现有算法相比，该算法能够提取出更准确的结果。

The experimental result of the algorithm which is proposed indicates that, it can extract more accurate results compared with existing.

参考文献：

Ramu, Gonapalli, and S. B. G. Thilak Babu. “Image Forgery Detection for High Resolution Images Using {SIFT} and {RANSAC} Algorithm.” 2017 2nd International Conference on Communication and Electronics Systems ({ICCES}), {IEEE}, 2017, doi:10.1109/cesys.2017.8321205.

完整源码下载地址：

http://page2.dfpan.com/fs/7lcaj2321d29016a504/

更多精彩文章请关注微信号：