首页 | 官方网站   微博 | 高级检索  
     

基于边缘特征的伪造图像盲检测算法
引用本文:李哲,郑江滨.基于边缘特征的伪造图像盲检测算法[J].西北工业大学学报,2009,27(5).
作者姓名:李哲  郑江滨
作者单位:西北工业大学,计算机学院,陕西,西安710072
基金项目:西北工业大学基础研究基金资助 
摘    要:文章针对伪造图像中常用的羽化操作,提出了一种基于边缘特征的检测伪造图像的方法.首先介绍了羽化操作模型,然后定义了象素的方向特征,并将梯度的概念应用到彩色图像中,最后结合象素的方向特征相容度和梯度方向角相容度,定位出图像的伪造区域.实验证明,该算法可以很好地检测出图像的真伪.

关 键 词:盲检测  伪造图像  羽化  边缘特征  梯度

A Novel Method of Blind Detection of Digitally Forged Image
Li Zhe,Zheng Jiangbin.A Novel Method of Blind Detection of Digitally Forged Image[J].Journal of Northwestern Polytechnical University,2009,27(5).
Authors:Li Zhe  Zheng Jiangbin
Abstract:Aim. In our opinion, all tampered images have one thing in common: feather operation. We propose a novel method of blind detection through discovering the traces of feather operation. Subsection 1.1 of the full paper introduces the model of feather operation. Subsection 1.2 briefs the well-known procedure of calculating the gradient of colored image. Subsections 1.3 and 1.4 do essentially three things: calculating the direction feature compatibility, calculating the gradient compatibility and locating the digitally forged region. The results of three experimental examples, given respectively in Figs. 1, 2 and 3 in the full paper, demonstrate preliminarily the effectiveness of our novel algorithm.
Keywords:blind detection  digitally forged image  direction feature compatibility  gradient compatibility
本文献已被 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号