结合面向对象方法的计算机视觉语义网络表示

The Representation of Semantic Nets Combining with Object Oriented Methodology in Computer Vision

本文从理论上分析一种新的计算机视觉关系模型表示方法:面向对象语义网络。这种表示方法减少了节点与语义联系的种类和数量,简化了表示形式。在面向对象语义网络的类里特别引进了属性的缺省值,在推理中引入缺省推理机制,这种方法符合人类思维方式;在类和对象中引进了操作,可以通过现有的知识推理新的知识;在属性中引进运动属性知识,使得这种语义网络更加有效地表达运动的场景。

The essay interprets in principle the representation of a new computer vision relationship model:Semantic Nets of Object Oriented Methodology.This representation reduces the variety and quantity of node and semantic re lationship,simplify its indication manner.It is matching the human's thinking method with introducing default value of attributesin particular in semantic nets of object oriented methodology and adopting default reasoning in reasoning mechanism.Adopting class and object in operation can help reason new knowledge through the current knowledge;adopting moving attributes in attributes can help semantic nets indicate the moving scene more effectively.