benchmarking for steganography by kernel fisher discriminant criterion

ISCAS OpenIR

	benchmarking for steganography by kernel fisher discriminant criterion
	Huang Wei; Zhao Xianfeng; Feng Dengguo; Sheng Rennong
	2012
会议名称	7th China International Conference on Information Security and Cryptography, Inscrypt 2011
会议录名称	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
页码	113-130
会议日期	November 30, 2011 - December 3, 2011
会议地点	Beijing, China
收录类别	EI
ISSN	0302-9743
ISBN	9783642347030
部门归属	(1) Institute of Software Chinese Academy of Sciences Beijing 100190 China; (2) State Key Laboratory of Information Security Institute of Information Engineering Chinese Academy of Sciences Beijing 100029 China; (3) Beijing Institute of Electronic Technology and Application Beijing 100091 China
摘要	In recent years, there have been many steganographic schemes designed by different technologies to enhance their security. And a benchmarking scheme is needed to measure which one is more detectable. In this paper, we propose a novel approach of benchmarking for steganography via Kernel Fisher Discriminant Criterion (KFDC), independent of the techniques in steganalysis. In KFDC, besides between-class variance resembles what Maximum Mean Discrepancy (MMD)merely concentrated on, within-class variance plays another important role. Experiments show that KFDC is qualified for the indication of the detectability of steganographic algorithms. Then, we use KFDC to illustrate detailed analysis on the security of JPEG and spatial steganographic algorithms. © 2012 Springer-Verlag Berlin Heidelberg.; In recent years, there have been many steganographic schemes designed by different technologies to enhance their security. And a benchmarking scheme is needed to measure which one is more detectable. In this paper, we propose a novel approach of benchmarking for steganography via Kernel Fisher Discriminant Criterion (KFDC), independent of the techniques in steganalysis. In KFDC, besides between-class variance resembles what Maximum Mean Discrepancy (MMD)merely concentrated on, within-class variance plays another important role. Experiments show that KFDC is qualified for the indication of the detectability of steganographic algorithms. Then, we use KFDC to illustrate detailed analysis on the security of JPEG and spatial steganographic algorithms. © 2012 Springer-Verlag Berlin Heidelberg.
关键词	Algorithms Cryptography Steganography
语种	英语
内容类型	会议论文
URI标识	http://ir.iscas.ac.cn/handle/311060/15855
专题	中国科学院软件研究所
推荐引用方式 GB/T 7714	Huang Wei,Zhao Xianfeng,Feng Dengguo,et al. benchmarking for steganography by kernel fisher discriminant criterion[C],2012:113-130.