图像数据库中一种基于OBFM的关联规则提取算法研究  

OBFM-based algorithm for extraction of correlative rules in image database

作　　者：庞淑侠[1] 袁占亭[1] 张秋余[1] 李明[1] 

机构地区：[1]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2005年第6期100-102,共3页Journal of Lanzhou University of Technology

基　　金：国家科技攻关项目(2001BA201A32)

摘　　要：针对传统的图像检索方法基本都是基于图像的固有特征而进行的,检索结果都存在较大误差,提出了一种OBFM(Ontology_BasedFeedbackModel)方法,首先得到基本相似图像库,然后采用DEDE原则进行特征元素的提取,并在此基础上进行关联规则提取.该方法大大提高了图像检索的准确率.Conventional image retrieval methods are essentially based on the intrinsic character of the image, so that there is larger error with the searching result. An OBFM（ontology-based feedback method） was presented and use first to get an image database with essentially similar images. Then the characteristic elements were extracted according to DEDE principle and, on this basis, the correlative rules were extracted. This method greatly improves the accuracy of image searching.

关 键 词：图像检索 本体 关联规则 

分 类 号：TP311.134[自动化与计算机技术—计算机软件与理论]