The Role of Lexical Ontology in Expanding the Semantic Textual Content of On-Line News Images

Using low-level features to support semantic search of images is a difficult task. As a result, textual content is used to provide semantic description or annotation of images. Such textual description of what we may call as 'surrounding text' is a value added features available in most web images particularly on-line news images. Most search engines used them as a feature to provide textual meaning of images. Relying on surrounding text alone, however, unable to provide support for semantic search that go beyond indexed terms. Lexical resources and ontology are potential sources to enhance searching for images. This paper discusses the use of WordNet and ConceptNet to enhance searching for on-line news images. This is further improved with named entity recognition (NER) technique to annotate important entities such as name if a person, location and organization among image searchers. Results show that lexical ontology has the capacity to semantically enhance the meanings of conventional bag of words index.