利用图像边缘变化特性寻找模糊处理痕迹

Detecting Blurred Traces based on Changing Feature of Image Edge

  • 摘要: 目前在图像认证技术中,一种不同于水印认证的盲检测技术发展迅速,这种认证技术不需要事先插入认证数据,只是根据图像本身的数据特征进行认证,其中对于图像处理痕迹的检测是盲检测的重要手段之一。本文提出了一种针对图像高斯模糊处理痕迹的检测算法,根据图像处理中高斯模糊的基本特性,首先对被检测原始图像采用不同的高斯半径进行二次模糊处理,然后用提出的算法来比较二次模糊后图像边缘的变化情况,算法可以从中发现原始图像的高斯模糊特性,从而发现原始图像中异常模糊处理的痕迹。仿真实验表明这种方法对于实现图像篡改的被动盲检测具有较好的应用价值。

     

    Abstract: In the domain of digital image forensics, a novel detecting approach called Blind Detection has been developing rapidly. Unlike Digital Watermarking, which needs watermarks be embedded into an image beforehand, Blind Detection technology is able to detect image based only on the intrinsic features of an image. Among which the detection of image manipulating traces is one of the crucial means in digital image forensics. An algorithm of detecting the trace of blur processing is proposed in this paper. Different Gaussian radii are utilized in re-blurring an original image, according to the fundamental of Gaussian Blur in image processing. Then changes of image edge after being blurred are compared to detect the different blur features in the original image, thereby traces of blur processing can be found with our proposed algorithm. The simulation results indicate that our proposed method is valuable in Passive Blind Image Forensics.

     

/

返回文章
返回