图像空域分析下的篡改检测算法

Blind Detection of Digital Forgery Image under Image Spatial Analysis

提出一种图像空域分析下的图像篡改检测算法。首先,利用GPAC分割出相似性区域;然后,利用双边滤波对相似性区域的灰度图进行去噪,并得到其噪声图像;最后,通过比较噪声图像的灰度值分布是否相同或噪声分布规律相似性程度来判断图像的真伪,并判断出其篡改方式。通过实验证明该算法能有效的检测出篡改图像。

A method about blind detection of digital forgery image under image spatial analysis is proposed. Firstly,the algorithm is using GPAC seg- ment similarities area, then, denoising similarities area of gray image based on bilateral filtering, and getting the noise image. Finally, the digital forgery could be detected through comparing the noise image distribution of grey value or the similarity of noise distribution shown by the statistics among the sim- ilarities area, and judge the way of tamper. Experiment results demonstrate the effectiveness of the algorithm in detecting the digital forgery image.