首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于颜色差分特性的图像篡改检测
引用本文:卢燕飞,牛佩佩,周骥.基于颜色差分特性的图像篡改检测[J].北方交通大学学报,2014(2):61-67.
作者姓名:卢燕飞  牛佩佩  周骥
作者单位:北京交通大学电子信息工程学院,北京100044
基金项目:国家自然科学基金资助项目(61071143)
摘    要:由于自然场景中目标物体表面的反射特性,以及图像采集过程中的CFA插值算法会使得图像三颜色分量之间具有相关性,特别是颜色分量的高频信息之间有近似相等的关系.拼接篡改操作会降低颜色高频分量的谱间相关性,同时也会给图像带来额外的高频信息.本文利用自然图像颜色通道高频分量之间的相关性,分析颜色差分图像中高频信息的分布情况来检测是否有拼接篡改操作.实验证明本文提取的高频特征分类效果较好,能够检测出图像的拼接篡改边缘.

关 键 词:CFA插值  相关性  高频  颜色差分  拼接

Digital image forensics based on the color differential characteristics
LU Yanfei,NIU Peipei,ZHOU Ji.Digital image forensics based on the color differential characteristics[J].Journal of Northern Jiaotong University,2014(2):61-67.
Authors:LU Yanfei  NIU Peipei  ZHOU Ji
Institution:(School of Electronics and Information Engineering, Beijing J iaotong University, Beijing 100044, China)
Abstract:Due to the reflective properties of the target object in natural scene and the CFA interpolationalgorithm of imaging process, there is a high correlation among red, green and blue channels. Especially, highfrequency of three color components are approximately equal. Splicing process will reducethe correlation of three color channels and introduce some additional high frequency. In this paper, wepresent a new method that uses the interchannel correlation to analyze the distribution of highfrequency in color difference image and detect the tampered regions. Experimental results show that theproposed method has a good performance, which can accurately locate the splicing edges in imageforgery.
Keywords:CFA interpolation  correlation  high-frequency  color difference  splicing
本文献已被 维普 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