面向多媒体文档的多通道检索系统  

A MULTI-CHANNEL RETRIEVAL SYSTEM FOR MULTIMEDIA DOCUMENTS

作　　者：汪燕云[1] 刘翔[2] 吴飞[2] 

机构地区：[1]杭州师范大学初等教育学院,浙江杭州310036 [2]浙江大学计算机学院,浙江杭州310027

出　　处：《计算机应用与软件》2009年第7期9-13,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(60603096)

摘　　要：提出和实现了一个面向多媒体文档的多通道(对应多种模态,如文本、图像、视频等)检索系统。系统定义了一个新的用来描述多媒体文档内容的框架,该框架不但提取出多媒体文档在各通道下的基于内容的底层特征,而且还记录下多媒体文档中不同多媒体对象间的链接关系。同时,提出一种基于图模型的交叉参照知识库,用来存储从链接关系中挖掘出的多媒体对象间的语义关系,通过一个有效的语义上下文分析算法,在检索过程中计算每个对象与查询的语义相似度。语义上下文分析算法不仅使得基于内容的多媒体信息检索中考虑了多媒体对象的语义信息,同时支持用户通过通道切换的方式进行相关反馈,提供了一种较为灵活的查询模式。实验表明在交叉参照知识库基础上,该系统还能够有效地提高基于内容的多媒体对象的检索性能(如内容覆盖率等)。A multi-channel, corresponding to multiple modals, e. g. text, image, video, etc. , retrieval system for multimedia documents is proposed and implemented. A novel framework model is defined in system to describe the contents of multimedia documents. In the framework, content-based bottom features in each channel extracted from multimedia documents and the links between multimedia objects of multimedia documents are both recorded. Moreover, a graph model-based cross-reference knowledge base is proposed to store the semantic relation between multimedia objects mined from link relations. A valid algorithm for semantic context analysis is proposed to calculate the semantic similarity between each object and query when retrieving. The algorithm not only endues content-based multimedia information retrieval method with semantic information of multimedia objects, but also provides a flexible query model, which enables a user to interact with the system through channel exchange and relevance feedback. It is demonstrated in the experiments that the retrieval performance （ e. g. the content coverage rate） for multimedia objects in the proposed multi-channel retrieval system is much better than that of a system with conventional content-based retrieval methods.

关 键 词：多媒体文档 多通道检索 语义上下文 内容覆盖率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP311.13[自动化与计算机技术—计算机科学与技术]