摘要
为更加准确地反映MPEG视频的内容,构造了一种将视频时间信息、运动特征信息以及DCT系数块能量关系信息等多种特征融合在一起的多特征数字签名算法,并以不增加MPEG编码复杂度的水印嵌入方式将数字签名嵌入到视频中,并实现了基于多特征的视频认证方案。认证实验结果证明,本方案对于如代码转换、重新压缩等常规处理较之单特征认证方法虚警概率要低,而对于如移除、替换等恶意攻击能做出有效判断,并能对篡改区域进行准确定位。
In order to more exactly reflect the MPEG video content, a multi-feature based digital signature by fusing the video temporal feature and motion feature and DCT coefficient feature was constructed, which was embedded into videos as watermarking without increasing the complexity of MPEG coding, and a new multi-feature based MPEG video au- thentication scheme was proposed. Experimental results show that the scheme is more robust than the single-feature one to the common processing such as code-switching and recompression, while sensitive to the malicious attacks such as erasure and insertion and replacement. It can further locate the corrupt areas.
出处
《通信学报》
EI
CSCD
北大核心
2008年第2期59-65,共7页
Journal on Communications
基金
国家重点基础研究发展计划(“973”计划)基金资助项目(2007CB310800)
湖北省自然科学基金资助项目(2007ABA402)~~
关键词
多特征
融合
数字签名
视频认证
multi-feature
fusion
digital signature
video authentication
