
2016 年 10 月 Journal on Communications October 2016
2016259-1
第 37 卷第 Z1 期 通 信 学 报 Vol.37
No.Z1
采用局部强度顺序模式的图像复制—粘贴篡改检测算法
林晶
1
,黄添强
2,3
,林玲鹏
2,3
,李小琛
1
(1. 福建师范大学数学与计算机科学学院,福建 福州 350007;2. 福建师范大学软件学院,福建 福州 350007;
3. 福建师范大学大数据分析与应用福建省高校工程研究中心,福建 福州 350007)
摘 要:复制—粘贴篡改是一种最简单而且常见的图像篡改方式。为了提高目前复制—粘贴篡改检测算法的顽健
性,提出一种基于局部强度顺序模式(LIOP,local intensity order pattern)的图像复制—粘贴篡改检测算法。首先,
提取待测图像的 LIOP 特征描述子,然后以特征描述子间的夹角余弦值作为相似性度量,根据最近邻与次近邻的
比值阈值寻找稳定的匹配点,最后计算匹配点对间的空间距离以移除误匹配点。实验结果表明,所提算法能够有
效检测并定位复制粘贴篡改位置,而且算法检测的准确率高,能够抵抗缩放、旋转、亮度变化以及高斯模糊、加
性高斯白噪声、JPEG 压缩等后期处理操作。
关键词:图像篡改检测;复制—粘贴篡改;特征描述;局部强度顺序模式
中图分类号:TP391 文献标识码:A
Detection of image copy-move forgery using
local intensity order pattern
LIN Jing
1
, HUANG Tian-qiang
2, 3
, LIN Ling-peng
2, 3
, LI Xiao-chen
1
(1.School of Mathematics and Computer Science, Fujian Normal University, Fuzhou 350007, China;
2. Faculty of Software, Fujian Normal University, Fuzhou 350007, China;
3. Fujian Provincial University Engineering Research Center of Big Data Analysis and Application, Fujian Normal University, Fuzhou 350007, China)
Abstract: Copy-move forgery was one of the most simple and common way of image manipulations. To improve the ro-
bustness of most existing copy-move forgery detections, a new method based on local intensity order pattern was pro-
posed. First, the LIOP feature descriptors were exacted from the inspected image. Then the angular cosine of feature de-
scriptors were used to measure the similarity, and the stable matching points were found according to the distance ratio
threshold of the nearest neighbor point to the second nearest neighbor. Finally, the space distance of the matching points
were calculated to remove the false matching points. Extensive experimental results were presented to confirm that the
proposed method is not only able to effectively identify and locate the altered area, but also have high accuracy and ro-
bust to scaling, rotation, brightness change and some post-processing, such as Gaussian blur, additive white Gaussian
noise and JPEG compression.
Key words: image tampering detection, copy-move forgery, feature description, local intensity order pattern
收稿日期:2016-09-01
通信作者:黄添强,fjhtq@fjnu.edu.cn
基金项目:国家自然科学基金资助项目(No.61070062,No.61502103) ;福建省高校产学合作科技重大基金资助项目
(No.2015H6007);福州市科技计划基金资助项目(No.2014-G-76);福建省高等学校新世纪优秀人才支持基金资助项目
(No.JAI1038);福建省科学厅 K 类基金资助项目(No.2011007);福建省教育厅 A 类基金资助项目(No.JA10064);福建师范大
学研究生教育改革研究基金资助项目(No.MY201414)
Foundation Items: The National Natural Science Foundation of China (No.61070062, No.61502103), Industry-University Coopera-
tion Major Projects in Fujian Province(No.2015H6007), Science and Technology Program of Fujian (No.2014-G-76), Program fo
ew Century Excellent Talents in University in Fujian Province(No. JAI1038), The Science and Technology Department of Fujian
Province K-Class Foundation Project (No.2011007), The Education Department of Fujian Province A-Class Foundation Project
(No.JA10064), The Graduate Education Reform Project of Fujian Normal University(No.MY201414)
doi:10.11959/j.issn.1000-436x.2016259