Video steganalysis method based on spatial-timeredundancy of statistics invisibility
一种基于时空冗余统计不可见性的视频隐写分析方法

The video steganography of the spread spectrum embedding and MSU are the two typical steganographic method , which can be resistant against compression and other attacks, in order to obtain the effective detection of the embedded secret information. According to the video spatial-time redundancy, this paper proposed a real-time video steganalysis method. Used a size of L+1 sliding window to obtain an estimate of the video frames, and extracted the corresponding DCT and Markov features, and used neural networks, support vector machines and other classification methods for video steganalysis. The results show that, according to the DCT and Markov features, the correct detection rate is higher. The support vector machines and time and spatial redundancy, etc can be used in the video steganalysis, which have great prospects.