基于时间和空间冗余的视频隐写分析方法

Video steganalysis method based on temporal and spatial redundancies

为了实现对视频隐写的有效检测,提出一种使用支持向量机(SVMs)的视频隐写盲检测方法。该方法利用时间和空间冗余,用帧间共谋的方式获取帧估计数据,提取视频帧的融合马尔科夫和DCT特征,构造SVMs分类器对待测视频进行检测,从而达到视频隐写检测的目的。实验结果表明,该方法能够准确地区分载密视频和原始视频。

In order to achieve effective video steganography detection, a video steganalysis scheme using support vector machines （SVMs） to blind detect video steganograms was proposed. It apply temporal and spatial redundancies obtain an estimate of the frame by interframe collusion and extract the merged Markov and discrete cosine transform （DCT） features to construct SVMs classifier to detect under test videos, so as to achieve the purpose of the video steganalysis. Results show that the our scheme accurately distinguish the cover and stego videos.