权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

RESEARCH AND DEVELOPMENT OF A NULTIMEDIA AND 4 DIMENSIONAL DATABASE SYSTEM

多媒体二维数据库系统的研发

基本信息

批准号：
07558158
负责人：
MAKINOUCHI Akifumi
金额：
$ 1.02万
依托单位：
KYUSHU UNIVERSITY
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (A)
财政年份：
1995
资助国家：
日本
起止时间：
1995 至 1997
项目状态：
已结题

项目摘要

Multimedia data such as videos and images are often represented and stored in their own file formats. Such data representation makes it difficult and time-consuming to retrieve them by contents. On the contrary, contents of multimedia data are modeled by spatio-temporal data base technologies in our system to accelerate multimedia data retrieval, because spatio-temporal data retrieval is expected to be much faster. From this point of view, we have studied the following.1.A data model for representing and storing spatio-temporal data is designed and implemented. In the model, spatio-temporal data is represented as a subspace in the space resulted from the Cartesian product of spatial, temporal, and non-spatio-temporal attributes. The subspace is represented by a simplicial complex. This representation guarantees that calculation of predicates and operators on spatio-temporal data is carried out in a unified way of combinatorial geometry.2.Spatio-temporal query language is studied by designing a spatio-temporal database and describing spatio-temporal queries. Spatio-temporal predicates are defined as well spatial and temporal predicates. Also, the problem of how spatio-temporal queries are described is studied. In addtion, the transformation of spatio-temporal queries into the queries in the internal representation is studied.3A spatio-temporal index is studied to retrieve them efficiently. The index is implemented based on R^<**>-tree. Spatio-temporal query processing is based on internal representation. This requires query processing on high dimensional spaces. Normalization method of such high dimensional space is studied to accelerate high dimensional spatial data.4.A graphical user interface for displaying spatio-temporal data is implemented using VRML and its browser. Using this interface, a user can understand what the high dimensional spatial data means visually.

像视频和图像这样的多媒体数据通常以它们自己的文件格式表示和存储。这种数据表示方式使得按内容检索它们变得困难且耗时。相反，我们的系统采用时空数据库技术对多媒体数据的内容进行建模，以加快多媒体数据的检索速度，因为时空数据的检索速度要快得多。从这个角度出发，我们研究了以下内容。设计并实现了一个表示和存储时空数据的数据模型。在该模型中，时空数据被表示为空间中由空间、时间和非时空属性笛卡尔积产生的子空间。子空间由一个简单复形表示。这种表示保证了谓词和运算符在时空数据上的计算以组合几何的统一方式进行。通过设计时空数据库和描述时空查询，研究了时空查询语言。时空谓词也定义为空间谓词和时间谓词。此外，本文还研究了如何描述时空查询的问题。此外，还研究了将时空查询转换为内部表示查询的问题。研究了3A时空索引，以实现对它们的高效检索。索引是基于R^<**>-tree实现的。时空查询处理基于内部表示。这需要在高维空间上进行查询处理。研究了这种高维空间的归一化方法，以加速高维空间数据的生成。利用VRML及其浏览器实现了显示时空数据的图形用户界面。使用这个界面，用户可以直观地理解高维空间数据的含义。