首页 | 本学科首页   官方微博 | 高级检索  
     

基于DCT系数统计特性的JPEG图像隐写分析
引用本文:徐志杰,蔡晓霞. 基于DCT系数统计特性的JPEG图像隐写分析[J]. 太赫兹科学与电子信息学报, 2010, 8(1): 79-83
作者姓名:徐志杰  蔡晓霞
作者单位:1. 汽车管理学院,干部轮训大队,安徽,蚌埠,233011
2. 电子工程学院,信息系,安徽,合肥,230037
摘    要:提出了一种基于离散余弦变换(DCT)系数统计特性的JPEG图像隐写分析算法。该算法在分析JPEG图像DCT系数的统计特性的基础上,提取了8维特征向量,通过最小二乘支持向量机(LS—SVM)分类器对待测图像进行分类,以达到检测载密图像的目的。算法实现简单,计算复杂度低。实验结果表明,该算法检测速度快,具有较高的检测正确率,能够实现针对各类JPEG图像信息隐写算法的有效检测。

关 键 词:隐写分析  DCT系数  特征向量  JPEG图像
收稿时间:2009-02-03
修稿时间:2009-11-02

Steganalysis for JPEG image based on statistical characteristics of DCT coefficients
XU Zhi-jie and CAI Xiao-xia. Steganalysis for JPEG image based on statistical characteristics of DCT coefficients[J]. Journal of Terahertz Science and Electronic Information Technology, 2010, 8(1): 79-83
Authors:XU Zhi-jie and CAI Xiao-xia
Affiliation:1.Cadre Rotational Training Brigade, Automotive Management Institute, Bengbu Anhui 233011, China; 2.Information Department, Electronic Engineering Institute, Hefei Anhui 230037, China)
Abstract:A steganalysis algorithm based on statistical characteristics of Discrete Cosine Transform (DCT) coefficients for Joint Photographic Experts Group(JPEG) image was proposed. The algorithm analyzed statistical characteristics of DCT coefficients, calculated a total of 8 dimensional feature vectors from each JPEG image, and classified cover and stego images using Least Squares Support Vector Machines(LS-SVM) classifier, therefore it could be applied to blind steganalysis. The algorithm could be implemented conveniently and computed easily. The experiments results showed that the algorithm had good performance in both detection rate and computation speed, and it was reliable in steganalysis for all kinds of JPEG images steganography.
Keywords:steganalysis  DCT coefficients  feature vectors  JPEG image
本文献已被 维普 万方数据 等数据库收录!
点击此处可从《太赫兹科学与电子信息学报》浏览原始摘要信息
点击此处可从《太赫兹科学与电子信息学报》下载免费的PDF全文
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