摘要
MPEG视频双压缩检测在视频取证领域中扮演着重要的角色。然而,在现有的双压缩检测算法中,多数算法只能检测不同比特率的双压缩视频,缺少一种有效的检测视频同比特率双压缩的办法。本文提出了一种基于宏块模式的视频同比特率双压缩检测算法。宏块模式由宏块的编码类型以及运动矢量组成,文中首先分析了宏块模式在多次使用相同比特率编码后的收敛性,之后选择利用支持向量机对其建立模型进行分类,判断待检测视频是否经过了同比特率的二次压缩。通过与现有的方法进行对比,实验结果表明本文提出的方法可以有效检测同比特率下的双压缩视频,并且同时适用于MPEG-2以及MPEG-4标准。
The detection of double MPEG compression plays an important role in video forensics. However, most existing methods could not effectively detect double MPEG compression with the same bitrates in constant bitrates coding. Thus, a novel method based on macro block mode (MBM) is proposed to detect double MPEG compression with the same bitrates. MBM is composed of macro block coding type and its motion vector. The convergence of MBM during repeated compression is firstly analyzed, then the support vector machine is used to model the MBM feature and to discriminate whether a video is doubly compressed with the same bitrates. Ex- periments indicate that the proposed method outperforms the existing methods and is robust to both MPEG-2 and MPEG-4 videos.
出处
《信息安全与通信保密》
2016年第2期113-116,122,共5页
Information Security and Communications Privacy
关键词
双压缩检测
同比特率
宏块模式
double compression
same bitrates
macro block mode