【作者】 郭慧慧；

【导师】 陈俊杰；

【作者基本信息】 太原理工大学 ， 计算机应用技术， 2010， 硕士

【摘要】 图像检索一直以来都是众多学者的研究内容之一,发展至今,大致可以分为四种类型,有基于内容的图像检索,有基于图像语义的检索,有基于反馈的信息检索技术,还有基于知识的信息检索系统。然而目前为止,并未有研究者将模糊理论应用到图像检索中,而本文实现了将模糊概念格理论应用于图像情感语义检索的过程,且由分析可知,这是可行的。由于图像所包含的情感是丰富的,不确定的,故将模糊形式概念分析理论应用于情感本体的构建也是合理的,从而更好的实现用户情感的检索。本文在同课题组的研究基础上查阅大量中外文献,基于MPEG-7的图像标准,提出了模糊情感本体框架;接着在模糊情感的基础上,将模糊概念格理论应用到情感本体库的构建中,设计了模糊情感概念格;并且在语义情感方面使用了基于VAD/PAD和OCC空间的情感分层内容,从而使得最终的情感检索能更好的实现用户的需求。最后在这些研究内容的基础上,提出了一个图像情感语义检索模型,并通过实验,对模型的可行性及有效性进行了验证。本文所做工作对情感本体库的研究做了有益的尝试。本文所做工作如下:(1)应用模糊形式概念分析方法构建了模糊情感概念格。该概念格是基于模糊形式概念分析方法的。模糊形式概念分析方法可以很好的用于情感本体的抽取,而抽取出的情感本体又可用于后续的研究。(2)设计了一个模糊情感匹配规则。这是一个可以用于模糊情感推理的匹配规则,算法所提取出的情感规则是模糊的,可以用于情感的模糊推理,从而得出一定的情感规则,用于后续的情感检索过程。(3)基于MPEG-7在情感方面的标准以及模糊概念格,构建了模糊情感本体框架。(4)基于模糊情感本体,设计了一个图像情感语义检索模型。该模型提供了情感语义在实际操作中的检索流程,为之后的程序设计提供基础。更多还原

【Abstract】 Image retrieval has been a hot topic for a great mount of researchers. Up to the present, there are about four kinds of image retrieve, which are content-based, image semantic-based, feedback-based and knowledge-based. However, no one use the fuzzy theory in the image searching. This paper basically implements it which is applying the concept lattice theory to the image retrieval process of emotional semantics, and the analysis shows that it is feasible.As the image contains a wealth of emotions and is uncertain, it is reasonable that the fuzzy Formal Concept Analysis theory is applied to build the emotion ontology. So that we can get a better search results for the user’s emotional retrieve.On the basis of our research group’s work and a large number of Chinese and foreign literature, the paper proposes a fuzzy ontology frame of emotion which is based on the standard of MPEG-7.Then, the paper applies the fuzzy concept lattice theory to the construction procedure of emotional ontology library, which is on the basis of the fuzzy affection, and designs a fuzzy emotion concept lattice. And the paper uses the layered emotion contents of VAD/PAD and OCC in the selecting of semantic affection space which can make the user be more satisfied with the final image search results.Finally, the paper presents a semantic image retrieval model on the basis of the previous studies and an experiment is completed to verify the feasibility and effectiveness of the model. And it is also a good attempt to the research of the emotional ontology library.The work we do list as follows:(1) The fuzzy emotion concept lattice by using the Fuzzy Formal Concept Analysis (FFCA) is constructed.The emotion concept lattice is based on the FFCA method. The FFCA method can be very good for the extraction of emotional ontology, and the extracted emotional ontology can also be used for subsequent studies(2) An algorithm of fuzzy emotional mapping rules is designed.This is a matching algorithm that can be used for the fuzzy emotional reasoning. The rules that be extracted by the algorithm are fuzzy and can be used for emotional fuzzy reasoning. Then the emotional mapping rules got by the reasoning process can be used for the subsequent retrieval process of emotion.(3) A fuzzy emotional ontology frame is built based on the relevant standard of MPEG-7 and fuzzy concept lattice.(4) A semantic retrieval model of image is designed on the basis of frame of fuzzy emotional ontology.The model provides the retrieval process in the actual operation and it can be a basis for the design of subsequent program.更多还原