SPAM数据挖掘Kim算法的隐写分析  

The Steganalysis of Kim Algorithm Based on SPAM Data Mining

作　　者：柳玲[1] 陈同孝 朱顺痣[1] 曹晨[1] 洪炜冬[1] 

机构地区：[1]厦门理工学院计算机科学与技术系,福建厦门361024 [2]台中科技大學资讯工程系

出　　处：《哈尔滨理工大学学报》2013年第4期31-36,共6页Journal of Harbin University of Science and Technology

基　　金：厦门市科技计划项目(3502Z20103036);台湾国科会计划项目(100-2221-E-024-014-MY2);厦门市科技计划高校创新项目(3502Z20123037)

摘　　要：隐写分析的目的主要是判断数字载体是否嵌有秘密信息.为了研究Kim等人提出的可逆数据隐藏算法的隐写性,用SPAM伪装分析器对其进行了隐写分析.实验中,选取RSP图库中5 000张大小为512×512的8位灰阶图像,对算法设置不同嵌入等级和采样间隔,以7种藏量来测试其抗SPAM检测的情况.最后,给出不同参数下算法抗SPAM检测变化的规律.实验结果表明,一般情况下,Kim等人提出的算法抗SPAM检测的平均错误率为0.88～21.00%,在嵌入等级较大时,抗SPAM检测的错误率可达30%以上,为以该算法作为载体来进行消息传递提供了参考.The steganalysis is to estimate if the digital carrier takes the secret messages. In order to study the steganalysis of Kim＇ s algorithm, SPAM camouflage analyzer was used to analyze Kim＇ s algorithm. In the experi- ment, 5,000 grayscale images with size 512 - 512 coming from RSP image database were firstly selected, 7 differ- ent embedding capacities were used against SPAM detection where different embedding level with various sub-sam- piing intervals under Kim algorithm were set. And then the rules of different parameters under Kim＇ s algorithm a- gainst SPAM detection were presented. Experimental results show that the average error rate of Kim＇ s algorithm a- gainst the SPAM detection is 0. 88% -21.00%, especially when embedding level is increasing, the average error rate can reach more than 30%, which provides academic references for Kim＇ s algorithm as a carrier to transmit message.

关 键 词：可逆数据隐藏 SPAM 隐写分析 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]