一个图像数据库检索系统的结构设计和快速检索方法  

Architecture Design and Fast Retrieving Method of An Image Database Retrieving System

作　　者：陈韶斌[1] 丁明跃[1] 周成平[1] 陈迎[1] 

机构地区：[1]华中科技大学图像识别与人工智能研究所

出　　处：《计算机与数字工程》2001年第3期34-40,64,共8页Computer & Digital Engineering

摘　　要：本文介绍了图像数据库区别于常规数据库的特点,并以数字化图书馆的图像数据库子系统Basestar为例,介绍了图像数据库检索系统的结构和检索方法。Basestar采用综合了颜色、形状、纹理三种特征矢量的综合性检索方法,本文详细介绍了这种方法,并论述了在此方法基础上,根据三角形不等式准则改进而成的快速检索方法的原理和实现。实验证明Basestar系统具有较好的检索效果,检索结果证明快速检索方法是有效的。In this paper, the speciality of an image database different from a general database is described. With the illustration of Basestar, the image database subsytem in digital library, the architecture and retrieving method of the image database retrieving system is stated. Basestar employs retrieving method synthesized with color feature vector, shape feature vector and texture feature vector. In this paper,this synthesized method is described in details,and the fast algorithm on the basis of this method, according to triangle inequality criterion, is also represented. Experiments shows that Basestar has a good retrieving effect and the effectivity of the fast retrieving method is proved by retrieving results.

关 键 词：图像数据库 特征矢量 快速检索 图像数据库检索系统 模式识别 

分 类 号：TP311.134[自动化与计算机技术—计算机软件与理论] TP391.41[自动化与计算机技术—计算机科学与技术]