基于自然语言处理的图像情感语义检索研究  

RESEARCH ON IMAGE EMOTIONAL SEMANTICS RETRIEVAL BASED ON NATURAL LANGUAGE PROCESSING

作　　者：于昕[1] 郭浩[1] 李海芳[1] 陈俊杰[1] 

机构地区：[1]太原理工大学计算机科学与技术学院,山西太原030024

出　　处：《计算机应用与软件》2014年第6期37-41,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(60970059;61070077;61170136);山西省自然科学基金项目(2010011020-2;2011011015-4);山西省教育厅高校科技项目(20121003);太原理工大学青年基金项目(2012L014)

摘　　要：自然语言问答在情感图像检索中并未广泛应用,对此进行了探索与尝试,设计和实现了一个以自然语言作为检索入口的情感图像检索系统,并给出系统架构及技术方案。研究过程中,引入自然语言处理技术对自然问句进行浅层语义分析,并建立了一个情感映射模型,实现了常用情感词语与该模型之间以及该模型与情感图像之间的映射。另外,还对系统的检索结果进行了评价。通过对系统的实现,为自然语言问答与情感图像检索的结合提供了一种新的思路与方法。Natural language question-answer has not been widely used in emotional image retrieval, on this issue we have explored and tried, and design and implement an emotional image retrieval system in which the natural language is used as a seareh entrance. The system architecture and technical solutions are also given in the article. In the process of the research, by introducing the natural language processing technique, we parse the shallow semantics of natural language questions and establish an emotional mapping model, which realises the mapping between the emotional words in common use and the model as well as between the model and the emotional images. Besides, we also evaluate the search results of the system. Through the implementation of the system, we provide a new train of thought and the method for the combination of natural language question-answer and emotion image retrieval.

关 键 词：自然语言处理 图像情感语义检索 情感模型映射 OCC模型PAD模型 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]