Abstract: Texture, recurrent number of object and the area of object are not taken into account in conventional low-dimension image association rules. So the relevant image knowledge can’t be mined out. The functions of 7D image association rules model 7D_AR proposed in this paper are more self-contained than conventional low-dimension image association rules, and can solve the above-mentioned problems better. Through concept generalization and deleting irrelevant dimension, various association rules less than 7D can be evolved from the 7D image association rules model 7D_AR.

Key words: Association rules model; Image mining; 7D_AR model

摘要： 传统的低维图像关联规则没有考虑纹理要素以及对象的重复出现次数、对象的面积大小，与此相关的图像知识无法挖掘出来。该文提出的七维图像关联规则模型7D_AR 功能比传统的低维图像关联规则更加完备，可以很好地解决以上问题。通过概念提升及删除无关维，由七维图像关联规则模型7D_AR 可以演化出维数小于七维的各种关联则。

关键词: 关联规则模型；图像挖掘；7D_AR 模型