作者: Wei Huang , Xianfeng Zhao , Dengguo Feng , Rennong Sheng
DOI: 10.1007/978-3-642-34704-7_10
关键词:
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 measure which one more detectable. this paper, we propose novel approach of for steganography via Kernel Fisher Discriminant Criterion (KFDC), independent the techniques in steganalysis. KFDC, besides between-class variance resembles what Maximum Mean Discrepancy (MMD)merely concentrated on, within-class plays another important role. Experiments show that KFDC qualified indication detectability algorithms. Then, use illustrate detailed analysis on security JPEG and spatial