使用混合DCT特征检测空域隐写术  

Spatial Steganography Detection Using Merged DCT Features

作　　者：张昊[1] 平西建[1] 

机构地区：[1]信息工程大学信息系统工程学院,河南郑州450002

出　　处：《信息工程大学学报》2012年第6期646-649,共4页Journal of Information Engineering University

基　　金：国家自然科学基金资助项目(60970142)

摘　　要：文章研究了一类流行的混合DCT特征对空域隐写术的检测能力。通过该特征对6种常见的空域隐写算法的隐写检测实验,得到以下新结论:①尽管该特征是针对JPEG隐写术设计的盲检测特征,它们也能有效检测常见的空域隐写术;②仅使用混合特征中的共生矩阵和Markov特征能够获得较混合特征更高的检测性能;③共生矩阵特征和校准技术的应用可以提升特征对产生分块效应的空域隐写术的检测能力。This paper focuses on the ability of a type of merged DCT steganalytic features to detect spatial domain steganography. Through testing the features with steganalysis experiments on six com- mon steganography algorithms, we have obtained the following new conclusions： （1） although the fea- tures are designed for detecting JPEG steganography, they can also detect the common spatial steg- anography effectively; （2）only using the co-occurrence matrix and Markov features can help us ac- quire higher detecting ability; （3） the co-occurrence matrix feature and calibration technique are use- ful for improving the features＇ detection ability on the algorithms which may produce the blockiness.

关 键 词：隐写术 隐写分析 混合DCT特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]