摘要
本文探讨计算机静态图像处理问题,提出两组六层的处理模型,即把图像内容分为高级和低级两组,各组再分别分为三层。计算语言学的主要任务是解决对低级内容实行自动化处理的问题,语料库语言学的主要任务是对高级内容实行人工切分与标注,以实行机助人标协同运作。知识本体在两组六层的处理模型中的作用有三:一是把人类知识形式化,二是供计算机处理图像时随时调用,三是对图像检索实行语义查询。
Nowadays it is no longer a challenge to compile a large size corpus of still images. The real difficulty lies in still image processing. This paper proposes a model of processing consisting of 6 layers divided into the high-level and low-level content groups. Computational linguistics focuses on the automatic processing of the low-level content,whereas corpus linguistics deals with the high-level content mainly manually through human segmentation and annotation. The bulk of the paper discusses and demonstrates the roles the OWL-based ontology plays in ( 1 ) the human segmentation and annotation of the high-level content,and ( 2 ) the assistance in the semantic retrieval of images from a large size image corpus.
出处
《当代语言学》
CSSCI
北大核心
2013年第2期214-229,252,共16页
Contemporary Linguistics
关键词
静态图像处理
知识本体
两组六层模型
机助人标
image corpus,low-level content,high-level content,ontology,image retrieval
