基于直方图统计分类隐写检测算法设计  

THE STEGANOGRAPHY DETECTION ALGORITHM DESIGN BASED ON HISTOGRAM STATISTICAL CLASSIFICATION

作　　者：邱志宏[1] 

机构地区：[1]柳州职业技术学院广西柳州545006

出　　处：《计算机应用与软件》2013年第6期159-161,165,共4页Computer Applications and Software

基　　金：广西教育厅科研课题(200103YB168,200911LX490)

摘　　要：为了提高隐写检测算法的检测性能,提出基于直方图统计分类的隐写检测算法。通过对图片信息的直方图特征参数的提取,使用构建基于人工神经网络的分类模式,实现对嵌入隐写信息的图片准确判定。详细分析基于直方图统计分类隐写算法的设计原理和过程,最后构建实验测试环境。测试结果表明,该隐写检测算法的检测成功率和误检率均优于Ezsteg0检测工具。In order to improve the detection performance of steganography detection algorithm, in this paper we put forward a steganography detection algorithm which is based on histogram statistical classification. Through extracting the histogram characteristic parameters of the image information and using the classification mode which is constructed based on artificial neural networks, this algorithm achieves accurate judgement on images embedded with steganography information. We analyse in detail the design, principle and process of this steganography algorithm, and at last construct the experimental test environment. Test results indicate that the detection success rate and the false alarm rate of the steganography detection algorithm designed in the paper are better than those of Ezstego detection tool.

关 键 词：隐写检测 直方图 神经网络 分类算法 特征 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]