放大连通数灵敏且健壮的视频隐写分析

A Robust Sensitive Video Steganalysis Scheme Base on MCN

针对DCT域的视频隐写,提出一种放大连通数(MCN,Magnifies Connected Numbers)的视频隐写分析新方法。在DCT域中隐写秘密消息时,秘密消息会不规则地分布到整个帧中。通过选取特殊的归一化系数的方法,被隐写帧的连通数会发生脉冲式增加,放大了隐写产生的影响,极大地提高了检测隐写的灵敏度。该方法能确定隐写位置、容量,且健壮性强,能抵抗噪声、滤波、剪切、旋转、分辨率重建、插入帧、删除帧等处理与攻击。

Based on video steganography of DCT domain,this paper presents a creative steganalysis scheme：MCN,i.e.Magnified Connected Number.Messages hidden in DCT domain will be unevenly distributed in the frames,selecting a specified value of the normalized coefficient,connected number ofthe stego-frame will has pulsing increase,steganography impact can be magnified,steganography detection sensitivity can be improved obviously.This scheme can detect location and capacity,It is robust,which can resist noise,filter,shear,rotation,resolution,inserting,deleting frame processing and attack.