图像隐写分析系统性能评价方法研究  

Research on Performance Evaluation for Steganalysis

作　　者：马秀莹[1] 

机构地区：[1]上海应用技术学院计算机科学与信息工程学院,上海201418

出　　处：《电信科学》2010年第S2期120-123,共4页Telecommunications Science

摘　　要：随着计算机取证技术的发展,有必要建立计算机取证工具的评价标准。在分析了常用的评价指标的缺点的基础上,提出采用KL距离比的相对度量模型评价隐写分析系统性能。用SαS模型描述图像的DCT系数分布,对掩密图像和原始图像分布之间的KL距离进行度量。利用该模型对常见的针对LSB的隐写分析方法进行了分析评价,同时研究了顺序嵌入和随机嵌入方式对度量的影响。实验表明,本文提出的模型能够有效地评价隐写分析系统性能。With the development of computer forensics,it is necessary to establish benchmark for computer forensic tools.Based on the analysis of common metrics,a novel metric represented by the ratio of KL divergence is proposed in the paper.Distribution of DCT coefficients is described by SαS model using which KL divergence between the distributions of cover and stego image is measured.Several LSB based steganalysis techniques are evaluated.Moreover,the effects of sequential and random LSB on evaluation are studied.Experimental results show that the metric given here can measure the performance of steganalysis schemes effectively.

关 键 词：手机取证 反取证技术 手机证据 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]