Enhancing Image Search Accuracy Using Semantic Annotation

Abstract

With the advent of the Internet as a medium of sharing resources like text, images, audio, video, documents etc., the Web users made it a way to interchange the knowledge as well. This behavior of the Web users over the Internet phenomenally turned the WWW into a huge repository of unstructured data inducing a need of the standard based representation of the data over the Internet. This need made itself more demanding because the searching or extraction of the knowledge from the unorganized data was becoming impossible, giving rise to the concept of Semantic Web. The use of XML for representing the data made the situation at the rest in the meanwhile. But, only the syntactical exploitation of the data could not help the situation because the process like searching of the images on the Web demanded the inclusion of the supplement of semantic structures into the list of standards. The Resource Description Framework (RDF) standard, a base technology of the Semantic Web, appeared as an intuitive solution of this problem as it employs the concept of annotation to describe the images and keeps all pieces of information pertaining to the images in the RDF documents, making the search process semantic rather than the traditional. In this thesis, the intent is to generate annotated RDF model for semantic retrieval of images using RDF editor for annotation. This RDF model is validated through online W3C RDF validation service. SPARQL, RDF query engine, is used to query the validated RDF model to check the efficiency of image search.