摘要
为了自动构建在外部参考框架下图像中对象间空间关系的描述,基于空间方向关系形式化模型四叉树直方图模型,采用基于四叉树直方图的模糊空间方向关系判断方法,建立了基于模糊规则的空间方向关系词、程度修饰词和距离关系词的选取方法,最终生成类似人的空间关系自然语言描述.人工图形与真实图片的实验结果表明该方法是适用的.
In linguistic scene descriptions, how to automatically build the description of spatial relations between objects is an important task of linguistic scene description. A method of fuzzy directional rela- tions that rely on the computation of quadtree histogram is introduced. And a system is constructed able to choose the suitable directional relations terms, the distance relations terms, a human-like linguistic de- scription of the spatial relations is generated as then. Excellent results from several synthetic and real im- age examples show that the applicability of this approach works well.
出处
《北京邮电大学学报》
EI
CAS
CSCD
北大核心
2012年第3期11-15,共5页
Journal of Beijing University of Posts and Telecommunications
基金
国家自然科学基金重大研究计划项目(90920006)
关键词
空间关系
外部参考框架
四叉树直方图
自然语言描述
模糊集
spatial relations
extrinsic reference framework
quadtree histogram
natural language de- scriptions
fuzzy set