一种人工模糊的伪造图像盲检测方法

针对模糊操作常常用来掩饰伪造拼接合成图像的拼接边缘但使边缘宽度增大的特点,提出了一种基于边缘宽度的伪造图像检测方法,该方法通过确定图像的边缘像素,计算边缘的宽度,去除较细边缘保留宽边缘等手段确定伪造数字图像的篡改区域。经实验验证,文中方法能有效地检测出伪造图像中经过模糊操作的边缘,从而确定伪造区域。

An Useful Blind Detection Method of Digital Forgery Image of Artificial Blur

Sections 1 through 3 of the full paper explain the blind detection method mentioned in the title,which we believe is novel and useful and whose core consists of: "Artificial blurring operation is generally used to conceal the combination edge of the image;blurring naturally brings along changes in the features of the edge.We analyzed the characteristics of the blurred edges and defined the width of edge.A novel method based on the edge width of blind detection is proposed to detect the forged digital images;the proposed method finds the edge pixels,calculates the variation of edge width,and eliminates original thin edges to detect the forgery areas." The results of experiments,presented in Figs 4 through 7,and their analysis demonstrate preliminarily that the proposed method can detect the artificially blurred edge of digital image except for special cases and is therefore useful for locating the forgery areas.