The Research On Image Annotation Applied In Image Retrieval

With the development of the technology of multimedia and internet, the automatic semantic image annotation has emerged as a hot topic where computer vision, machine learning and natural language understanding have already been applied in recent years. However, the result of image annotation can not be satisfied and the image annotation is also a challenged problem in the domain of image retrieval. In this paper, by analyzing the traditional state of art image annotation models, we use the relevant knowledge of graph based learning in machine learning, fuzzy set and information fusion to solve the problem in image annotation. We address the issue of image annotation and a method of image annotation with two bag level generators, an approach to image annotation with semantic fusion, an image annotation scheme using vague fusion and a scheme of thesaurus search-based image annotation are proposed in this paper. At the same time, the technology of vague set and thesaurus are first applied in this domain, a novel thesaurus search based image retrieval framework is proposed in this paper. The validity of these methods are all performed by experiments. These proposed methods in this paper broad the prospect for the development of image retrieval.