基于图像背景噪声相关性的篡改检测  

Image Forgery Detection Using Correlation of Background Noise

作　　者：洪拥筠 刘本永[1,2] 

机构地区：[1]贵州大学大数据与信息工程学院,贵州贵阳550025 [2]贵州大学智能信息处理研究所,贵州贵阳550025

出　　处：《贵州大学学报（自然科学版）》2015年第3期93-96,118,共5页Journal of Guizhou University:Natural Sciences

基　　金：国家自然科学基金项目资助(60862003);科技部国际合作项目资助(2009DFR10530);贵州省工业科技攻关项目资助(黔科合GY字(2010)3054号)

摘　　要：图像拼接是一种简单而常用的数字图像篡改操作,本文利用拼接篡改图像背景噪声特性的不一致性,提出一种基于图像背景噪声相关性的拼接图像篡改检测方法。该方法首先利用小波变换对待测图像去噪,再将待测图像与去噪后的图像相减后得到背景噪声图像,最后根据背景噪声图像中子块与周围邻域子块的相关系数大小来鉴别拼接篡改图像并定位篡改位置。实验结果表明该方法能对拼接篡改图像较为准确地鉴别和定位。Generally,the characteristics of background noise are inconsistent in a splicing tamper image,which is common in digital tampering images. In this paper,a method for splicing tamper image forgeries detection,using correlation of background noise,was proposed. First,a background noise image was gotten by subtracting denoised image from the image which needs to detect. Then according to the correlation of each sub-block and that of neighborhood block in the background noise image,we could identify and locate tampered region. The experimental results show that this method can accurately identify splicing tampered image and locate tampered region.

关 键 词：图像篡改检测 背景噪声 小波变换 相关系数 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]