摘要： 视频信息处理的关键是视频信息的结构化，视频除了有基本层次结构之外，还有隐藏其中的视频结构语法和结构语义。该文提出了一种视频结构挖掘的概念框架和视频结构挖掘的系统框架，在概念框架中对视频结构挖掘相关概念给出了明确定义和界定；视频结构挖掘框架主要包括：视频基本层次结构挖掘，视频结构语法挖掘和视频结构语义挖掘。讨论了视频结构模式和知识的具体应用，包括指导视频的组织与管理、实现基于内容的个性视频推荐和改善视频摘要系统。

关键词: 视频结构挖掘, 视频基本层次结构, 视频结构语法, 视频结构语义

Abstract: Besides the basic hierarchical structure which consists of four kinds of components, namely frame, shot, scene, and video program, a video has structure syntactic and structure semantic hiding behind the video basic structure. This paper proposes a video structure mining framework, which includes conception framework and mining system framework. In the conception framework, the definitions of the relational concepts are given; in the mining system framework, the main contents have video basic hierarchical structure mining, video structure syntactic mining, and video structure semantic mining. Three applications of video structure patterns and knowledge are discussed, which include: guiding video management, indexing and retrieval; achieving personalized recommender systems of video commodity based on content; and improving video summarization systems.

Key words: Video structure mining, Video basic hierarchical structure, Video structure syntactic, Video structure semantic