基于概念本体的视频内容分析框架  

Video Content Analysis Framework Based on Concept Ontology

作　　者：张良[1] 周长胜[1] 

机构地区：[1]北京信息科技大学计算中心,北京100192

出　　处：《计算机技术与发展》2011年第12期116-119,123,共5页Computer Technology and Development

基　　金：北京市属高等学校人才强教计划资助项目(PHR201008447)

摘　　要：为解决视频内容分析过程中对视频组织结构划分的问题,建立了一个基于概念本体的视频分析模型,将视频以概念本体的方式进行划分,归纳出视频中四种典型的概念:类别概念、对象概念、属性概念以及情感概念。论述了各类概念之间的关系:实例关系、属性关系、归类关系、组成关系。采用人工定义规则与学习方法相结合的方法,实现对视频概念本体的识别和划分。以概念本体为中心划分视频结构,符合人类认知事物过程和认知规律,便于将底层特征向量与高层语义进行映射、组织和处理。In order to solve the video structure problem of semantic analysis,created a video content analysis framework based on concept ontology. The video is divided by the concept ontology, it summarized the concept of four typical video： category concept, object con- cepts, property concepts and emotional concepts. It also discusses the relationship between various concepts：instance-of, attribute-of,kind -of,part-of. The video structured by concept ontology consistent with human cognitive processes and cognitive rules of things. The concept ontology as the center can establish good contact between the underlying feature vector and high-level semantic,to the easy mapping,organization and processing.

关 键 词：视频 内容分析 概念本体 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]