| Ontology and folksonomy are both popular methods of information organization in the Internet.The purpose is to make it easier for users to manage information and obtain information.Folksonomy is a popular classification behavior,its biggest advantage is the universality.Ontology is a specification of shared information,it is an important part of the semantic web.The combination of the them is an organization of information,which can complement each other and give full play to their respective advantages to display network information resources.This paper analyzes the network image library and presents the information of the image library.First of all,this paper analyzes the related theories of folksonomy and ontology.By comparing the advantages and disadvantages of them.This paper analyzes the feasibility of applying the two to the image field.Extract concepts by analyzing the collection of tags in the flickr dataset.Diagnose the relationships between concepts by analyzing the visual features of the image,including symbiotic relations and inclusion relations.Secondly,the complete scheme of visual ontology construction is given,including: discovery of concepts,search of concept relations,construction of concept hierarchy,etc.And the corresponding evaluation criteria.Design CML(Concept Model Training)algorithms to find examples of each concept.That is,the number of images contained in each concept.DAG(directed acyclic graphs)are drawn from the mined relationship to reflect the relationships between the concepts.It is used to organize the basic information of the image library.Finally,the visual ontology is implemented through software.Detailed steps include: class creation,attribute addition,class and class instance description.Get ontology description language OWL(Ontology Web Language).It is used to build a shared web ontology.It Implements the sharing of network ontology among users.The application of ontology structure in fuzzy query and label recognition is elaborated. |
PDF Full Text Request