摘要
针对目前视频检索中低层特征与高层语义间存在"鸿沟"的问题,提出一种基于本体的视频检索方法,着重介绍本体框架的构造和以本体映射为核心的视频检索算法.实验结果表明该方法能够显著提高检索的查准率和查全率.
Aimed at the problem of the existing gap between low-level features and high-level semantic at present, a video retrieval approach based on ontology is proposed in this paper. A detailed discussion is made of the construction of the ontology framework and the video retrieval algorithm which focus on ontology mapping. Experiment results show that this method improved markedly in terms of retrieval precision and recall rate.
出处
《西南大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第11期119-124,共6页
Journal of Southwest University(Natural Science Edition)
关键词
视频本体
本体框架
视频检索
本体映射
video ontology
ontology infrastructure
video retrieval
ontology mapping
