面向内容网络环境的视频内容标识方案

Video content identification scheme in content oriented network environment

视频指纹技术是当前解决海量视频数据标识问题的一种有效方案。首先阐述了内容网络的研究现状以及存在的主要问题,尤其在内容标识方面的问题;然后从内容网络中的视频内容标识需求出发,提出视频内容标识方案的设计原则,并深入分析已有方案的原理和缺点;最后在解析视频指纹技术概念与原理之后,设计并实现一种基于视频指纹技术的视频内容标识方案,其特点为:将时域和频域两种类型的视频指纹技术进行融合,在缩短了视频标识生成时间的同时,提高了整体系统的顽健性。

Video fmgerprinting technology is a video identification technology proposed for the identification of massive video data. Firstly, the research status of the content network and the main problems that exist were elaborated, especially the problems of content identification. Then design principles of the video content identification scheme were proposed from the video content identification requirements in the content network, and the principles and disadvantages of the existing solutions were analyzed in depth. Finally, a video content identification scheme based on video fingerprinting technology was designed and implemented after analyzing the concept and principle of video fingerprinting technology. The feature was： the two types of video fingerprinting technology in the time domain and frequency domain were integrated. While shortening the time for generating video tags, it also improved the robustness of the overall system.