摘要
通过对视频帧序列时空特性的分析,采用"局部趋同,全局异化"的策略,提出了一种联合时域和空域SIFT点特征的特征提取方法。实验表明,基于该特征的同源视频检测方法对于一定的视频变化具有较好的鲁棒性和检测精度。
Video features extraction is the important part of detection of coderivative video .SIFT features are local features with excellent performance. By anglicizing the temporal and spatial characteristics of video frames sequences SIFT feature, this paper takes the ＂local convergence, global alienation＂ strategy, and proposes a feature extraction method of spatiotemporal SIFY characteristics. Experiments of detection of coderivative video based on the method show that the proposed co-features has better robustness against certain changes of video sequences ,and also has a better detection accuracy than some existing methods.
出处
《电子技术应用》
北大核心
2012年第3期130-133,共4页
Application of Electronic Technique
基金
国家"863"计划资助项目(2011AA010603)
关键词
同源视频检测
视频帧序列时空特性
联合时空SIFT特征
detection of coderivative video
spatiotemporal characteristics of video frames sequences
spatiotemporal SIFT char-acteristics
