【作者】 孙春华；

【导师】 王命延；

【作者基本信息】 南昌大学 ， 计算机软件与理论， 2006， 硕士

【摘要】 数字视频在广播、教育和其他多媒体应用领域发挥着越来越重要的作用。随着应用的发展，视频数据量也急剧增加，如何高效管理这些海量视频数据使用户能够快速获取所需的视频数据也越来越受到人们的重视。本文在分析研究现有一些基于内容和注释的视频分析、处理和检索技术的基础上，结合MPEG-7标准的新框架，提出一种自动化和人机交互的手段进行视频编目的方法。主要的研究工作如下： 1．视频结构化处理技术：基于视频层次化模型，对通用视频的结构化处理技术进行研究，提出了一种基于特定视频的领域知识来自动化实现视频结构化处理的方法。 2 视频内容描述：本文在研究MPEG-7多媒体描述接口标准的基础上，从视频数据的特性出发，综合考虑视频的各种特征，包括视觉特征、语义特征、时空结构，将视频描述分为视觉和语义两个层次，以自动化和半自动的人机交互模式建立基于MPEG-7标准的视频内容描述模型。 3．计算机辅助视频编目系统框架设计：基于J2EE平台采用三层架构设计。系统能够实现对视频进行分析、索引和基于WEB的视频检索功能。更多还原

【Abstract】 Digital video plays a more and more important role nowadays in the areas of broadcast, education and other multimedia applications. With the development of applications, the amount of video is increasing rapidly. People show great concern about how to manage efficiently immense amount of video to facilitate users to access video content they want to get. In this paper, on the base of studying earlier approaches of content-based and annotation-based video analysis, processing and retrieval, we describe a system for automatic and interactive video indexing. Main works in this thesis is as follows.1.Video analysis and processing. Based on hierarchical video model, study the technology of analysis and processing of generic video model. Present an approach of automatically processing of video structure in specific domain.2. Description of video content. Based on Mpeg-7 standard, integrating a variety of features, include visual features, semantic features, and temporal and spatial structure, construct the descriptions of low-level and high-level features of video content by human computer interaction methods.3. Design of the architecture of the System of Computer Aided Video cataloging. Design three tiered architecture based on the platform of J2EE to realize video analysis, indexing, and web-based retrieval.更多还原