基于内容的Flash数据库检索算法研究

发布时间：2018-04-25 17:30

本文选题：Flash动画 + 基于内容　；参考：《山东师范大学》2012年硕士论文

【摘要】：信息时代的到来，使得网络成为最大的多媒体教育资源库。Flash动画是近十年来发展起来的一种基于矢量图形的多媒体教育资源，具有交互功能、制作简单、表现能力强、视觉感染力大、便于网络传播等特点，在教学中深受广大师生的喜爱。在多媒体领域，图像、声音、视频等媒体形式都涌现出了很多专业的搜索系统，为广大网络用户提供了极大的便利。面对日益丰富的Flash动画资源，如何快速、准确的检索到需要的资源成为教育工作者研究的重点。Flash动画作为一种新的多媒体形式出现较晚，目前已经成型的Flash网络搜索引擎还很少，并且都是采用关键词搜索，利用它们从网上检索教学和学习所需的Flash动画资源的效率往往不高。因此，建立一个基于内容的Flash数据库检索系统，对于教育教学工作是非常有研究意义的。本文在研究基于内容的Flash检索的基础上，建立了一个基于内容的Flash数据库检索系统。该检索系统以Flash的总体内容特征为基本检索，以Flash的逻辑场景、视觉场景和组成元素的内容特征为高级检索。用户可以根据自己的需求发出检索请求,由系统根据检索条件计算数据库中各记录的相似度，将满足条件的Flash以Gif动画的形式作为检索结果提供给用户。本文会在已经做好的检索系统的基础上，重点着眼于改善Flash索引数据库的结构，完善检索系统的算法，以此来提高检索系统的检索速度和检索效率。本论文的主要工作如下： 1.本文深入研究了多媒体数据库的结构，根据Flash动画的内容特征描述，将每个Flash的所有内容特征定义为数据库的各个字段并将其存储为特征文件。数据库的结构借鉴了倒排文档的思想对特征文件进行倒排文档索引，并且对数据库进行了编码，提高了系统的检索速度、优化了系统的检索结果。 2.检索算法是本篇论文的核心内容，本文提出了一种利用文本检索技术进行基于内容的Flash检索的新方法，检索系统参考布尔模型的精确匹配思想来设计检索算法。用户输入的查询条件与Flash索引数据库中的字段完全匹配时，相似度定义为1，即能够检索到用户需要的Flash资源；否则相似度定义为0，检索不到用户需要的资源。 3.由于索引数据库用到了特殊的倒排文档方式来存储记录，而且基于内容的Flash检索系统采用了结构化查询方式来设计系统的检索算法，，于是为了优化检索系统的检索速度，检索系统引用了数组的思想来存储各个模块的检索结果。 4.系统在为用户输出结果之前，检索系统将结果记录集放入缓存，缓存的使用缩短了用户在输出页面时等待翻页的时间。本文旨在为用户提供快速、专业、准确的Flash动画资源检索服务。初步实验结果表明，系统检索结果的准确率较高，数据库结构的重新设计以及索引和缓存等的使用也使系统检索的速度明显提高。
[Abstract]:With the arrival of the information age, the network becomes the largest multimedia educational resource bank. Flash animation is a kind of multimedia educational resource based on vector graphics developed in the past ten years. It has interactive function, simple production and strong performance. Visual appeal, easy to network transmission and other characteristics, deeply loved by teachers and students in teaching. In the multimedia field, image, sound, video and other media forms have emerged a lot of professional search systems, for the vast number of network users to provide great convenience. In the face of the increasingly abundant Flash animation resources, how to quickly and accurately retrieve the needed resources has become the focus of educators' research. Flash animation, as a new multimedia form, has emerged relatively late. At present, there are few Flash web search engines, and all of them use keyword search. The efficiency of searching Flash animation resources for teaching and learning from the Internet is often not high. Therefore, the establishment of a content-based Flash database retrieval system is of great significance for education and teaching. Based on the research of content-based Flash retrieval, a content-based Flash database retrieval system is established in this paper. The retrieval system is based on the overall content feature of Flash and advanced retrieval based on the logical scene, visual scene and content feature of component elements of Flash. Users can send out retrieval requests according to their own requirements. The system calculates the similarity of each record in the database according to the retrieval conditions. The Flash that meets the conditions is provided to the users in the form of Gif animation as retrieval results. Based on the existing retrieval system, this paper focuses on improving the structure of the Flash index database and the algorithm of the retrieval system, so as to improve the retrieval speed and efficiency of the retrieval system. The main work of this thesis is as follows: 1. In this paper, the structure of multimedia database is deeply studied. According to the content feature description of Flash animation, all the content features of each Flash are defined as each field of the database and stored as feature files. The structure of the database uses the idea of inverted documents for reference to index the feature files and encodes the database to improve the retrieval speed of the system and optimize the retrieval results of the system. 2. Retrieval algorithm is the core of this paper. This paper presents a new method of content-based Flash retrieval using text retrieval technology. The retrieval system designs the retrieval algorithm with reference to the exact matching idea of Boolean model. When the query criteria input by the user and the fields in the Flash index database are exactly matched, the similarity is defined as 1, that is, the Flash resources needed by the user can be retrieved; otherwise, the similarity is defined as 0, and the resources needed by the user cannot be retrieved. 3. Because the index database uses the special inverted document way to store the records, and the content based Flash retrieval system adopts the structured query method to design the retrieval algorithm of the system, so in order to optimize the retrieval speed of the retrieval system, The retrieval system references the idea of array to store the retrieval results of each module. 4. Before the system outputs the result for the user, the retrieval system puts the result record set into the cache, and the use of the cache shortens the time for the user to wait for the page to turn when the page is output. This article aims to provide users with fast, professional, accurate Flash animation resources retrieval services. The preliminary experimental results show that the accuracy of the retrieval results is high, the database structure redesign and the use of index and cache also improve the speed of the system retrieval.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP317.4;TP311.13

【参考文献】