Research On Image Retrieval Based On Formal Concept Analysis

With the popularization of multimedia devices and the rapid development of the Internet, the digital image is growing at an alarming rate. In face of the growing digital images, how to retrieve the relevant images we demand in the database quickly and efficiently have become an issue people concerned.The current image retrieval tools are based on the text, the content and the semantic. The image retrieval based on the context and the content is the most common method. In order to improve the efficiency of image retrieval, the paper proposes a technique of image retrieval based on formal concept analysis. As a powerful tool for data analysis, rule extraction and knowledge discovery, FCA (Formal Concept Analysis) combines data effectively by using the concept lattice by which we can find the implicit knowledge of the concept inside. Recently the formal concept analysis has been widely used in machine learning, data mining and information retrieval and other fields. For improving the efficiency of image retrieval, this paper proposes image retrieval technology based on the formal concept analysis.Firstly, this paper studies the current situation of the domestic and international image retrieval, analyzes the practical application of formal concept analysis, and then put forward the image retrieval method based on formal concept analysis. This paper makes improvement mainly on the computing of the image similarity.Secondly, this paper gives the theoretical foundation of formal concept analysis, and introduces the construction method of the formal context and concept lattice, which is explained by simple examples in detail. This paper also makes summary of the concept lattice in the application of the mainstream.Then the paper introduces several current technologies of image retrieval which have been made brief description on the advantages and disadvantages. CBIR (content based image retrieval) can be divided into two parts:the extraction of image features (color features, texture features, shape feature) and the calculation of image similarity.Finally, Based on the theory of form concept analysis the paper gives the definition of the image context, generates the concept lattice of the image and presents the related knowledge of graph matching at the same time. After generating the concept lattice based on the formal context of images, the Hasse diagram corresponding with the concept lattice can be generated and the images which are similar to an example image also be found using the lattice matching method. The concept lattice similarity calculation methods are used to graph matching. Recently the formal concept analysis (mainly the concept of grid) has been widely applied to the fields of the machine learning, the data mining and the retrieval information calculation of the similarity of the image features. This paper searches 120 images randomly on the internet, retrieves the images by using the methods of formal concept analysis and verify the validity of FCA in image retrieval.