Semantic approach to image retrieval using statistical models based on a lexical ontology

Abstract

The increasing amount of digital images available on the Internet has made searching, browsing, and organizing such resources a major challenge. This paper proposes a semantic approach to text-based image retrieval of manually annotated digital images. The approach uses statistical models based on Semantic DNA (SDNA) extracted from the structure of a lexical ontology called OntoRo. The approach involves three main techniques: (a) SDNA extraction, (b) word sense disambiguation using statistical models based on the extracted SDNA, and (c) applying semantic similarity measures using SDNA. The experiments performed show that the proposed approach retrieves images based on their conceptual meaning rather than the use of specific keywords in their annotations. © Springer-Verlag 2010.