摘要： 网络视频信息包括视频内容和视频描述信息。鉴于效率问题，网络视频监管主要考虑视频描述信息。结合本体论和成熟的Java字符处理规则，提出基于Ontology的网络视频描述信息抽取方法。建立网络视频领域内的本体模型，自定义启发式抽取规则，并应用Java编程将该方法应用于某部委国家级视频监管项目中。实验结果表明，该方法在抽取的准确率和覆盖率上取得了较好的效果。

关键词: Ontology模型, 网络视频监管, Web信息抽取

Abstract: Internet video monitoring is mainly concerned with the textual video information based on the consideration of the efficient problem, although the video content information is very important. In this paper, a regular rule processing regular for the extraction problem based on the ontology theory is formally defined. An ontology model called Ontology Model of Internet Video Domain(IVD-OM) is designed for the video information, and an implementation by using Java program is developed. The proposed approach to a real-world application of a national department is applied. Experimental results show that the proposed approach is effective and can accurately extract the video information from the Web.

Key words: Ontology model, Internet video monitoring, Web information extraction

中图分类号: 

• TP391