联合小波和马尔可夫特征的数字图像来源认证  

Source digital image identification combined wavelet and Markov features

作　　者：范馨月[1] 李永强[1] 周非[1] 

机构地区：[1]重庆邮电大学移动通信技术重庆市重点实验室,重庆400065

出　　处：《计算机应用研究》2014年第6期1918-1920,共3页Application Research of Computers

基　　金：重庆市基础与前沿研究计划资助项目(cstc2013jcyjA40034);重庆市教委科学技术研究资助项目(KJ130528);重庆市科委重点实验室专项经费资助项目

摘　　要：针对数字图像来源判别准确率不高的问题,提出了基于模式噪声的图像来源认证方法。该算法首先利用小波降噪算法从已拍摄图像中提取模式噪声,然后求取模式噪声的小波和马尔可夫特征,最后采用支持向量机(support vector machine,SVM)分类方法,利用噪声特征组合的分类器进行图像来源分类,并对分类结果进行定性和定量分析。实验结果表明,该方法对不同品牌和同一品牌不同型号的相机都有较高的判别准确率。In order to improve the accuracy of source identification of digital image,this paper put forward the image source authentication method based on pattern noise.Firstly,the algorithm extracted the noise which had taken from the image,using the wavelet de-noising algorithm.Then it calculated the features of wavelet and Markov,and finally used the support vector machine,by the combination of the noise characteristics as a classifier to indentify source image.At the same time,it analysed the classified results qualitatively and quantitatively.The experimental results demonstrate that in allusion to different brand and different models of the same brand,the method has a high accuracy.

关 键 词：模式噪声 小波特征 马尔可夫特征 小波降噪 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]