作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 图形图像处理 • 上一篇    下一篇

基于色彩一致性的图像模糊篡改检测

周治平,胡成燕,黄浩   

  1. (江南大学物联网工程学院,江苏 无锡 214122)
  • 收稿日期:2014-12-12 出版日期:2016-01-15 发布日期:2016-01-15
  • 作者简介:周治平(1962-),男,教授、博士,主研方向为图像处理;胡成燕、黄浩,硕士研究生。

Image Blur Forgery Detection Based on Color Consistency

ZHOU Zhiping,HU Chengyan,HUANG Hao   

  1. (School of Internet of Things Engineering,Jiangnan University,Wuxi,Jiangsu 214122,China)
  • Received:2014-12-12 Online:2016-01-15 Published:2016-01-15

摘要: 分析数字图像中的局部色彩相关性及模糊篡改操作对图像色彩一致性的影响,提出一种基于改进色彩特征的数字图像模糊篡改检测方法。重新定义色调变化率和异常色调率,通过提取图像中的色调变化和异常色调特征,描述模糊操作对图像局部色彩一致性的破坏程度,将特 征输入支持向量机进行训练和分类,并对图像中的模糊篡改区域进行定位检测。实验结果表明,该方法可快速检测出图像中的模糊操作,准确定位模糊篡改区域,且对图像的旋转和JPEG压缩操作具有一定的鲁棒性。

关键词: 数字图像取证, 色调变化, 异常色调, 色彩一致性, 模糊篡改, 鲁棒性

Abstract: This paper analyzes the local color correlation and the influence of blur forgery operation on image color consistency,and proposes a new method based on the improvement of the color features to detect the blur tamper of digital image.By redefining the normal of hue bias rate and abnormal hues rate.It extracts the hue bias and abnormal hue features to describe the destruction of the local color consistency introduced by imaging pipeline which is caused by blurring operation.These features are fed to the Support Vector Machine(SVM) to train and classify,and it realizes position detection for image blur forgery regions.Experimental results show that the proposed method has a good performance in detecting blur forgery operation,it can accurately locate the regions of blur forgery,and this method also has a certain robustness of the rotating and JPEG compression operation in tampered regions.

Key words: digital image forensics, hue change, abnormal hue, color consistency, blur forgery, robustness

中图分类号: