摘要
在实真场景中,在载体失配(CSM,cover source mismatch)条件下降低虚警率是隐写分析的一个巨大挑战,提出了一种新的模型来处理该问题。该方法由来源分类器首先判断图像的来源,继而利用相关来源图像训练而成的隐写分类器判断待测图像是否为载密。在这个过程中,通过对模型参数的调节减小虚警率。实验结果表明,这种方法可以在较大准确率的前提下最小化虚警率。
In the real world, reducing false positive rates in the case of cover source mismatch(CSM) was a big challenge for steganalysis. A novel model was proposed to solve the problem. The proposed method determines the image-acquiring source firstly by a source detector and then detecting the steg images in each source with a steganalyzer trained for this source. The false positive rate was reduced by solving a parameter model. The experimental results show that this novel method can reach lower false positive rates for larger true positive rates.
作者
杨培韬
张卫明
俞能海
YANG Pei-tao ZHANG Wei-ming YU Neng-hai(CAS Key Laboratory of Electromagnetic Space Information, University of Science and Technology of China, Hefei 230001, China)
出处
《通信学报》
EI
CSCD
北大核心
2016年第12期165-170,共6页
Journal on Communications
基金
国家自然科学基金资助项目(No.61572452
No.61502007
No.U1636201)
中国博士后科学基金资助项目(No.2015M582015)
中国科学院战略性先导专项基金资助项目(No.XDA06030601)~~
关键词
虚警率
失配
隐写分析
最小化虚警模型
false positive
mismatch
steganalysis
minimum false positive model